├── lib ├── model │ ├── __init__.py │ ├── nms │ │ ├── __init__.py │ │ ├── _ext │ │ │ ├── __init__.py │ │ │ └── nms │ │ │ │ └── __init__.py │ │ ├── .gitignore │ │ ├── make.sh │ │ ├── src │ │ │ ├── nms_cuda_kernel.h │ │ │ ├── nms_cuda.h │ │ │ └── nms_cuda_kernel.cu │ │ ├── nms_gpu.py │ │ ├── nms_wrapper.py │ │ ├── build.py │ │ ├── nms_cpu.py │ │ └── nms_kernel.cu │ ├── rpn │ │ ├── __init__.py │ │ ├── generate_anchors.py │ │ ├── rpn.py │ │ └── proposal_layer.py │ ├── utils │ │ ├── __init__.py │ │ ├── .gitignore │ │ ├── times_b.ttf │ │ ├── blob.py │ │ ├── logger.py │ │ ├── viz_hand_obj.py │ │ ├── bbox.pyx │ │ └── matching.py │ ├── faster_rcnn │ │ ├── __init__.py │ │ ├── vgg16.py │ │ └── faster_rcnn.py │ ├── roi_align │ │ ├── __init__.py │ │ ├── _ext │ │ │ ├── __init__.py │ │ │ └── roi_align │ │ │ │ └── __init__.py │ │ ├── functions │ │ │ ├── __init__.py │ │ │ └── roi_align.py │ │ ├── modules │ │ │ ├── __init__.py │ │ │ └── roi_align.py │ │ ├── make.sh │ │ ├── src │ │ │ ├── roi_align.h │ │ │ ├── roi_align_cuda.h │ │ │ ├── roi_align_kernel.h │ │ │ ├── roi_align_cuda.c │ │ │ ├── roi_align.c │ │ │ └── roi_align_kernel.cu │ │ └── build.py │ ├── roi_crop │ │ ├── __init__.py │ │ ├── _ext │ │ │ ├── __init__.py │ │ │ ├── crop_resize │ │ │ │ └── __init__.py │ │ │ └── roi_crop │ │ │ │ └── __init__.py │ │ ├── modules │ │ │ ├── __init__.py │ │ │ └── roi_crop.py │ │ ├── functions │ │ │ ├── __init__.py │ │ │ ├── roi_crop.py │ │ │ ├── crop_resize.py │ │ │ └── gridgen.py │ │ ├── make.sh │ │ ├── src │ │ │ ├── roi_crop_cuda.h │ │ │ ├── roi_crop.h │ │ │ ├── roi_crop_cuda_kernel.h │ │ │ └── roi_crop_cuda.c │ │ └── build.py │ ├── roi_pooling │ │ ├── __init__.py │ │ ├── _ext │ │ │ ├── __init__.py │ │ │ └── roi_pooling │ │ │ │ └── __init__.py │ │ ├── functions │ │ │ ├── __init__.py │ │ │ └── roi_pool.py │ │ ├── modules │ │ │ ├── __init__.py │ │ │ └── roi_pool.py │ │ ├── src │ │ │ ├── roi_pooling.h │ │ │ ├── roi_pooling_cuda.h │ │ │ ├── roi_pooling_kernel.h │ │ │ ├── roi_pooling_cuda.c │ │ │ └── roi_pooling.c │ │ └── build.py │ ├── roi_layers │ │ ├── nms.py │ │ ├── __init__.py │ │ ├── roi_pool.py │ │ └── roi_align.py │ ├── csrc │ │ ├── vision.cpp │ │ ├── cpu │ │ │ ├── vision.h │ │ │ └── nms_cpu.cpp │ │ ├── nms.h │ │ ├── ROIPool.h │ │ ├── ROIAlign.h │ │ └── cuda │ │ │ ├── vision.h │ │ │ ├── nms.cu │ │ │ └── ROIPool_cuda.cu │ └── extension_layers │ │ └── extension_layers.py ├── pycocotools │ ├── __init__.py │ ├── UPSTREAM_REV │ ├── license.txt │ ├── maskApi.h │ ├── mask.py │ └── maskApi.c ├── datasets │ ├── ._voc_eval.py │ ├── ._pascal_voc.py │ ├── __init__.py │ ├── VOCdevkit-matlab-wrapper │ │ ├── get_voc_opts.m │ │ ├── xVOCap.m │ │ └── voc_eval.m │ ├── ds_utils.py │ ├── tools │ │ └── mcg_munge.py │ ├── factory.py │ ├── vg_eval.py │ └── voc_eval_raw.py ├── roi_data_layer │ ├── __init__.py │ ├── minibatch.py │ └── roidb.py └── setup.py ├── assets ├── method.png └── boardgame_848_sU8S98MT1Mo_00013957.png ├── requirements.txt ├── test.md ├── cfgs ├── vgg16.yml ├── res50.yml ├── res101.yml ├── res101try.yml └── res101_ls.yml ├── _init_paths.py ├── LICENSE └── .gitignore /lib/model/__init__.py: -------------------------------------------------------------------------------- 1 | -------------------------------------------------------------------------------- /lib/model/nms/__init__.py: -------------------------------------------------------------------------------- 1 | -------------------------------------------------------------------------------- /lib/model/rpn/__init__.py: -------------------------------------------------------------------------------- 1 | -------------------------------------------------------------------------------- /lib/model/utils/__init__.py: -------------------------------------------------------------------------------- 1 | -------------------------------------------------------------------------------- /lib/model/faster_rcnn/__init__.py: -------------------------------------------------------------------------------- 1 | -------------------------------------------------------------------------------- /lib/model/nms/_ext/__init__.py: -------------------------------------------------------------------------------- 1 | -------------------------------------------------------------------------------- /lib/model/roi_align/__init__.py: -------------------------------------------------------------------------------- 1 | -------------------------------------------------------------------------------- /lib/model/roi_crop/__init__.py: -------------------------------------------------------------------------------- 1 | -------------------------------------------------------------------------------- /lib/model/roi_pooling/__init__.py: -------------------------------------------------------------------------------- 1 | -------------------------------------------------------------------------------- /lib/model/roi_align/_ext/__init__.py: -------------------------------------------------------------------------------- 1 | -------------------------------------------------------------------------------- /lib/model/roi_crop/_ext/__init__.py: -------------------------------------------------------------------------------- 1 | -------------------------------------------------------------------------------- /lib/model/roi_crop/modules/__init__.py: -------------------------------------------------------------------------------- 1 | -------------------------------------------------------------------------------- /lib/model/roi_pooling/_ext/__init__.py: -------------------------------------------------------------------------------- 1 | -------------------------------------------------------------------------------- /lib/model/roi_align/functions/__init__.py: -------------------------------------------------------------------------------- 1 | -------------------------------------------------------------------------------- /lib/model/roi_align/modules/__init__.py: -------------------------------------------------------------------------------- 1 | -------------------------------------------------------------------------------- /lib/model/roi_crop/functions/__init__.py: -------------------------------------------------------------------------------- 1 | -------------------------------------------------------------------------------- /lib/model/roi_pooling/functions/__init__.py: -------------------------------------------------------------------------------- 1 | -------------------------------------------------------------------------------- /lib/model/roi_pooling/modules/__init__.py: -------------------------------------------------------------------------------- 1 | -------------------------------------------------------------------------------- /lib/model/nms/.gitignore: -------------------------------------------------------------------------------- 1 | *.c 2 | *.cpp 3 | *.so 4 | -------------------------------------------------------------------------------- /lib/model/utils/.gitignore: -------------------------------------------------------------------------------- 1 | *.c 2 | *.cpp 3 | *.so 4 | -------------------------------------------------------------------------------- /lib/pycocotools/__init__.py: -------------------------------------------------------------------------------- 1 | __author__ = 'tylin' 2 | -------------------------------------------------------------------------------- /assets/method.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ddshan/hand_object_detector/HEAD/assets/method.png -------------------------------------------------------------------------------- /lib/pycocotools/UPSTREAM_REV: -------------------------------------------------------------------------------- 1 | https://github.com/pdollar/coco/commit/3ac47c77ebd5a1ed4254a98b7fbf2ef4765a3574 2 | -------------------------------------------------------------------------------- /lib/datasets/._voc_eval.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ddshan/hand_object_detector/HEAD/lib/datasets/._voc_eval.py -------------------------------------------------------------------------------- /lib/model/utils/times_b.ttf: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ddshan/hand_object_detector/HEAD/lib/model/utils/times_b.ttf -------------------------------------------------------------------------------- /lib/datasets/._pascal_voc.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ddshan/hand_object_detector/HEAD/lib/datasets/._pascal_voc.py -------------------------------------------------------------------------------- /assets/boardgame_848_sU8S98MT1Mo_00013957.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ddshan/hand_object_detector/HEAD/assets/boardgame_848_sU8S98MT1Mo_00013957.png -------------------------------------------------------------------------------- /requirements.txt: -------------------------------------------------------------------------------- 1 | scipy 2 | pillow>=7.1.0 3 | easydict 4 | cython 5 | cffi 6 | opencv-python 7 | msgpack 8 | matplotlib 9 | pyyaml 10 | tensorboardX 11 | tqdm 12 | jinja2 13 | pascal_voc_writer 14 | -------------------------------------------------------------------------------- /lib/model/roi_pooling/src/roi_pooling.h: -------------------------------------------------------------------------------- 1 | int roi_pooling_forward(int pooled_height, int pooled_width, float spatial_scale, 2 | THFloatTensor * features, THFloatTensor * rois, THFloatTensor * output); -------------------------------------------------------------------------------- /lib/model/roi_layers/nms.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) Facebook, Inc. and its affiliates. All Rights Reserved. 2 | # from ._utils import _C 3 | from model import _C 4 | 5 | nms = _C.nms 6 | # nms.__doc__ = """ 7 | # This function performs Non-maximum suppresion""" 8 | -------------------------------------------------------------------------------- /lib/model/nms/make.sh: -------------------------------------------------------------------------------- 1 | #!/usr/bin/env bash 2 | 3 | # CUDA_PATH=/usr/local/cuda/ 4 | 5 | cd src 6 | echo "Compiling stnm kernels by nvcc..." 7 | nvcc -c -o nms_cuda_kernel.cu.o nms_cuda_kernel.cu -x cu -Xcompiler -fPIC -arch=sm_52 8 | 9 | cd ../ 10 | python build.py 11 | -------------------------------------------------------------------------------- /lib/model/nms/src/nms_cuda_kernel.h: -------------------------------------------------------------------------------- 1 | #ifdef __cplusplus 2 | extern "C" { 3 | #endif 4 | 5 | void nms_cuda_compute(int* keep_out, int *num_out, float* boxes_host, int boxes_num, 6 | int boxes_dim, float nms_overlap_thresh); 7 | 8 | #ifdef __cplusplus 9 | } 10 | #endif 11 | -------------------------------------------------------------------------------- /lib/model/roi_align/make.sh: -------------------------------------------------------------------------------- 1 | #!/usr/bin/env bash 2 | 3 | CUDA_PATH=/usr/local/cuda/ 4 | 5 | cd src 6 | echo "Compiling my_lib kernels by nvcc..." 7 | nvcc -c -o roi_align_kernel.cu.o roi_align_kernel.cu -x cu -Xcompiler -fPIC -arch=sm_52 8 | 9 | cd ../ 10 | python build.py 11 | -------------------------------------------------------------------------------- /test.md: -------------------------------------------------------------------------------- 1 | data/ 2 | ├── cache_handobj_100K 3 | ├── cache_handobj_100K_ego 4 | ├── cache_handobj_ego 5 | ├── pretrained_model 6 | ├── VOCdevkit2007_handobj_100K 7 | │ └── VOC2007 8 | │ ├── Annotations 9 | │ ├── ImageSets 10 | │ │ └── Main 11 | │ └── JPEGImages -------------------------------------------------------------------------------- /lib/datasets/__init__.py: -------------------------------------------------------------------------------- 1 | # -------------------------------------------------------- 2 | # Fast R-CNN 3 | # Copyright (c) 2015 Microsoft 4 | # Licensed under The MIT License [see LICENSE for details] 5 | # Written by Ross Girshick 6 | # -------------------------------------------------------- 7 | -------------------------------------------------------------------------------- /lib/model/roi_crop/make.sh: -------------------------------------------------------------------------------- 1 | #!/usr/bin/env bash 2 | 3 | CUDA_PATH=/usr/local/cuda/ 4 | 5 | cd src 6 | echo "Compiling my_lib kernels by nvcc..." 7 | nvcc -c -o roi_crop_cuda_kernel.cu.o roi_crop_cuda_kernel.cu -x cu -Xcompiler -fPIC -arch=sm_52 8 | 9 | cd ../ 10 | python build.py 11 | -------------------------------------------------------------------------------- /lib/roi_data_layer/__init__.py: -------------------------------------------------------------------------------- 1 | # -------------------------------------------------------- 2 | # Fast R-CNN 3 | # Copyright (c) 2015 Microsoft 4 | # Licensed under The MIT License [see LICENSE for details] 5 | # Written by Ross Girshick 6 | # -------------------------------------------------------- 7 | -------------------------------------------------------------------------------- /lib/model/nms/src/nms_cuda.h: -------------------------------------------------------------------------------- 1 | // int nms_cuda(THCudaTensor *keep_out, THCudaTensor *num_out, 2 | // THCudaTensor *boxes_host, THCudaTensor *nms_overlap_thresh); 3 | 4 | int nms_cuda(THCudaIntTensor *keep_out, THCudaTensor *boxes_host, 5 | THCudaIntTensor *num_out, float nms_overlap_thresh); 6 | -------------------------------------------------------------------------------- /lib/datasets/VOCdevkit-matlab-wrapper/get_voc_opts.m: -------------------------------------------------------------------------------- 1 | function VOCopts = get_voc_opts(path) 2 | 3 | tmp = pwd; 4 | cd(path); 5 | try 6 | addpath('VOCcode'); 7 | VOCinit; 8 | catch 9 | rmpath('VOCcode'); 10 | cd(tmp); 11 | error(sprintf('VOCcode directory not found under %s', path)); 12 | end 13 | rmpath('VOCcode'); 14 | cd(tmp); 15 | -------------------------------------------------------------------------------- /lib/datasets/VOCdevkit-matlab-wrapper/xVOCap.m: -------------------------------------------------------------------------------- 1 | function ap = xVOCap(rec,prec) 2 | % From the PASCAL VOC 2011 devkit 3 | 4 | mrec=[0 ; rec ; 1]; 5 | mpre=[0 ; prec ; 0]; 6 | for i=numel(mpre)-1:-1:1 7 | mpre(i)=max(mpre(i),mpre(i+1)); 8 | end 9 | i=find(mrec(2:end)~=mrec(1:end-1))+1; 10 | ap=sum((mrec(i)-mrec(i-1)).*mpre(i)); 11 | -------------------------------------------------------------------------------- /lib/model/roi_crop/modules/roi_crop.py: -------------------------------------------------------------------------------- 1 | from torch.nn.modules.module import Module 2 | from ..functions.roi_crop import RoICropFunction 3 | 4 | class _RoICrop(Module): 5 | def __init__(self, layout = 'BHWD'): 6 | super(_RoICrop, self).__init__() 7 | def forward(self, input1, input2): 8 | return RoICropFunction()(input1, input2) 9 | -------------------------------------------------------------------------------- /cfgs/vgg16.yml: -------------------------------------------------------------------------------- 1 | EXP_DIR: vgg16 2 | TRAIN: 3 | HAS_RPN: True 4 | BBOX_NORMALIZE_TARGETS_PRECOMPUTED: True 5 | RPN_POSITIVE_OVERLAP: 0.7 6 | RPN_BATCHSIZE: 256 7 | PROPOSAL_METHOD: gt 8 | BG_THRESH_LO: 0.0 9 | BATCH_SIZE: 256 10 | LEARNING_RATE: 0.01 11 | TEST: 12 | HAS_RPN: True 13 | POOLING_MODE: align 14 | CROP_RESIZE_WITH_MAX_POOL: False 15 | -------------------------------------------------------------------------------- /lib/model/roi_layers/__init__.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) Facebook, Inc. and its affiliates. All Rights Reserved. 2 | import torch 3 | from .nms import nms 4 | from .roi_align import ROIAlign 5 | from .roi_align import roi_align 6 | from .roi_pool import ROIPool 7 | from .roi_pool import roi_pool 8 | 9 | __all__ = ["nms", "roi_align", "ROIAlign", "roi_pool", "ROIPool"] 10 | -------------------------------------------------------------------------------- /lib/model/nms/nms_gpu.py: -------------------------------------------------------------------------------- 1 | from __future__ import absolute_import 2 | import torch 3 | import numpy as np 4 | from ._ext import nms 5 | import pdb 6 | 7 | def nms_gpu(dets, thresh): 8 | keep = dets.new(dets.size(0), 1).zero_().int() 9 | num_out = dets.new(1).zero_().int() 10 | nms.nms_cuda(keep, dets, num_out, thresh) 11 | keep = keep[:num_out[0]] 12 | return keep 13 | -------------------------------------------------------------------------------- /_init_paths.py: -------------------------------------------------------------------------------- 1 | import os.path as osp 2 | import sys 3 | 4 | def add_path(path): 5 | if path not in sys.path: 6 | sys.path.insert(0, path) 7 | 8 | this_dir = osp.dirname(__file__) 9 | 10 | # Add lib to PYTHONPATH 11 | lib_path = osp.join(this_dir, 'lib') 12 | add_path(lib_path) 13 | 14 | coco_path = osp.join(this_dir, 'data', 'coco', 'PythonAPI') 15 | add_path(coco_path) 16 | -------------------------------------------------------------------------------- /lib/model/roi_align/src/roi_align.h: -------------------------------------------------------------------------------- 1 | int roi_align_forward(int aligned_height, int aligned_width, float spatial_scale, 2 | THFloatTensor * features, THFloatTensor * rois, THFloatTensor * output); 3 | 4 | int roi_align_backward(int aligned_height, int aligned_width, float spatial_scale, 5 | THFloatTensor * top_grad, THFloatTensor * rois, THFloatTensor * bottom_grad); 6 | -------------------------------------------------------------------------------- /lib/model/roi_crop/_ext/crop_resize/__init__.py: -------------------------------------------------------------------------------- 1 | 2 | from torch.utils.ffi import _wrap_function 3 | from ._crop_resize import lib as _lib, ffi as _ffi 4 | 5 | __all__ = [] 6 | def _import_symbols(locals): 7 | for symbol in dir(_lib): 8 | fn = getattr(_lib, symbol) 9 | locals[symbol] = _wrap_function(fn, _ffi) 10 | __all__.append(symbol) 11 | 12 | _import_symbols(locals()) 13 | -------------------------------------------------------------------------------- /lib/model/roi_align/src/roi_align_cuda.h: -------------------------------------------------------------------------------- 1 | int roi_align_forward_cuda(int aligned_height, int aligned_width, float spatial_scale, 2 | THCudaTensor * features, THCudaTensor * rois, THCudaTensor * output); 3 | 4 | int roi_align_backward_cuda(int aligned_height, int aligned_width, float spatial_scale, 5 | THCudaTensor * top_grad, THCudaTensor * rois, THCudaTensor * bottom_grad); 6 | -------------------------------------------------------------------------------- /cfgs/res50.yml: -------------------------------------------------------------------------------- 1 | EXP_DIR: res50 2 | TRAIN: 3 | HAS_RPN: True 4 | # IMS_PER_BATCH: 1 5 | BBOX_NORMALIZE_TARGETS_PRECOMPUTED: True 6 | RPN_POSITIVE_OVERLAP: 0.7 7 | RPN_BATCHSIZE: 256 8 | PROPOSAL_METHOD: gt 9 | BG_THRESH_LO: 0.0 10 | DISPLAY: 20 11 | BATCH_SIZE: 256 12 | WEIGHT_DECAY: 0.0001 13 | DOUBLE_BIAS: False 14 | SNAPSHOT_PREFIX: res50_faster_rcnn 15 | TEST: 16 | HAS_RPN: True 17 | POOLING_MODE: crop 18 | -------------------------------------------------------------------------------- /cfgs/res101.yml: -------------------------------------------------------------------------------- 1 | EXP_DIR: res101 2 | TRAIN: 3 | HAS_RPN: True 4 | BBOX_NORMALIZE_TARGETS_PRECOMPUTED: True 5 | RPN_POSITIVE_OVERLAP: 0.7 6 | RPN_BATCHSIZE: 256 7 | PROPOSAL_METHOD: gt 8 | BG_THRESH_LO: 0.0 9 | DISPLAY: 20 10 | BATCH_SIZE: 128 11 | WEIGHT_DECAY: 0.0001 12 | DOUBLE_BIAS: False 13 | LEARNING_RATE: 0.001 14 | TEST: 15 | HAS_RPN: True 16 | POOLING_SIZE: 7 17 | POOLING_MODE: align 18 | CROP_RESIZE_WITH_MAX_POOL: False 19 | -------------------------------------------------------------------------------- /cfgs/res101try.yml: -------------------------------------------------------------------------------- 1 | EXP_DIR: res101 2 | TRAIN: 3 | HAS_RPN: True 4 | BBOX_NORMALIZE_TARGETS_PRECOMPUTED: True 5 | RPN_POSITIVE_OVERLAP: 0.7 6 | RPN_BATCHSIZE: 256 7 | PROPOSAL_METHOD: gt 8 | BG_THRESH_LO: 0.0 9 | DISPLAY: 20 10 | BATCH_SIZE: 128 11 | WEIGHT_DECAY: 0.0001 12 | DOUBLE_BIAS: False 13 | LEARNING_RATE: 0.001 14 | TEST: 15 | HAS_RPN: True 16 | POOLING_SIZE: 7 17 | POOLING_MODE: align 18 | CROP_RESIZE_WITH_MAX_POOL: False 19 | -------------------------------------------------------------------------------- /lib/model/roi_pooling/src/roi_pooling_cuda.h: -------------------------------------------------------------------------------- 1 | int roi_pooling_forward_cuda(int pooled_height, int pooled_width, float spatial_scale, 2 | THCudaTensor * features, THCudaTensor * rois, THCudaTensor * output, THCudaIntTensor * argmax); 3 | 4 | int roi_pooling_backward_cuda(int pooled_height, int pooled_width, float spatial_scale, 5 | THCudaTensor * top_grad, THCudaTensor * rois, THCudaTensor * bottom_grad, THCudaIntTensor * argmax); -------------------------------------------------------------------------------- /lib/model/nms/_ext/nms/__init__.py: -------------------------------------------------------------------------------- 1 | 2 | from torch.utils.ffi import _wrap_function 3 | from ._nms import lib as _lib, ffi as _ffi 4 | 5 | __all__ = [] 6 | def _import_symbols(locals): 7 | for symbol in dir(_lib): 8 | fn = getattr(_lib, symbol) 9 | if callable(fn): 10 | locals[symbol] = _wrap_function(fn, _ffi) 11 | else: 12 | locals[symbol] = fn 13 | __all__.append(symbol) 14 | 15 | _import_symbols(locals()) 16 | -------------------------------------------------------------------------------- /lib/model/roi_crop/_ext/roi_crop/__init__.py: -------------------------------------------------------------------------------- 1 | 2 | from torch.utils.ffi import _wrap_function 3 | from ._roi_crop import lib as _lib, ffi as _ffi 4 | 5 | __all__ = [] 6 | def _import_symbols(locals): 7 | for symbol in dir(_lib): 8 | fn = getattr(_lib, symbol) 9 | if callable(fn): 10 | locals[symbol] = _wrap_function(fn, _ffi) 11 | else: 12 | locals[symbol] = fn 13 | __all__.append(symbol) 14 | 15 | _import_symbols(locals()) 16 | -------------------------------------------------------------------------------- /lib/model/roi_align/_ext/roi_align/__init__.py: -------------------------------------------------------------------------------- 1 | 2 | from torch.utils.ffi import _wrap_function 3 | from ._roi_align import lib as _lib, ffi as _ffi 4 | 5 | __all__ = [] 6 | def _import_symbols(locals): 7 | for symbol in dir(_lib): 8 | fn = getattr(_lib, symbol) 9 | if callable(fn): 10 | locals[symbol] = _wrap_function(fn, _ffi) 11 | else: 12 | locals[symbol] = fn 13 | __all__.append(symbol) 14 | 15 | _import_symbols(locals()) 16 | -------------------------------------------------------------------------------- /lib/model/roi_pooling/_ext/roi_pooling/__init__.py: -------------------------------------------------------------------------------- 1 | 2 | from torch.utils.ffi import _wrap_function 3 | from ._roi_pooling import lib as _lib, ffi as _ffi 4 | 5 | __all__ = [] 6 | def _import_symbols(locals): 7 | for symbol in dir(_lib): 8 | fn = getattr(_lib, symbol) 9 | if callable(fn): 10 | locals[symbol] = _wrap_function(fn, _ffi) 11 | else: 12 | locals[symbol] = fn 13 | __all__.append(symbol) 14 | 15 | _import_symbols(locals()) 16 | -------------------------------------------------------------------------------- /lib/model/roi_crop/src/roi_crop_cuda.h: -------------------------------------------------------------------------------- 1 | // Bilinear sampling is done in BHWD (coalescing is not obvious in BDHW) 2 | // we assume BHWD format in inputImages 3 | // we assume BHW(YX) format on grids 4 | 5 | int BilinearSamplerBHWD_updateOutput_cuda(THCudaTensor *inputImages, THCudaTensor *grids, THCudaTensor *output); 6 | 7 | int BilinearSamplerBHWD_updateGradInput_cuda(THCudaTensor *inputImages, THCudaTensor *grids, THCudaTensor *gradInputImages, 8 | THCudaTensor *gradGrids, THCudaTensor *gradOutput); 9 | -------------------------------------------------------------------------------- /cfgs/res101_ls.yml: -------------------------------------------------------------------------------- 1 | EXP_DIR: res101 2 | TRAIN: 3 | HAS_RPN: True 4 | BBOX_NORMALIZE_TARGETS_PRECOMPUTED: True 5 | RPN_POSITIVE_OVERLAP: 0.7 6 | RPN_BATCHSIZE: 256 7 | PROPOSAL_METHOD: gt 8 | BG_THRESH_LO: 0.0 9 | DISPLAY: 20 10 | BATCH_SIZE: 128 11 | WEIGHT_DECAY: 0.0001 12 | SCALES: [800] 13 | DOUBLE_BIAS: False 14 | LEARNING_RATE: 0.001 15 | TEST: 16 | HAS_RPN: True 17 | SCALES: [800] 18 | MAX_SIZE: 1200 19 | RPN_POST_NMS_TOP_N: 1000 20 | POOLING_SIZE: 7 21 | POOLING_MODE: align 22 | CROP_RESIZE_WITH_MAX_POOL: False 23 | -------------------------------------------------------------------------------- /lib/model/csrc/vision.cpp: -------------------------------------------------------------------------------- 1 | // Copyright (c) Facebook, Inc. and its affiliates. All Rights Reserved. 2 | #include "nms.h" 3 | #include "ROIAlign.h" 4 | #include "ROIPool.h" 5 | 6 | 7 | PYBIND11_MODULE(TORCH_EXTENSION_NAME, m) { 8 | m.def("nms", &nms, "non-maximum suppression"); 9 | m.def("roi_align_forward", &ROIAlign_forward, "ROIAlign_forward"); 10 | m.def("roi_align_backward", &ROIAlign_backward, "ROIAlign_backward"); 11 | m.def("roi_pool_forward", &ROIPool_forward, "ROIPool_forward"); 12 | m.def("roi_pool_backward", &ROIPool_backward, "ROIPool_backward"); 13 | } 14 | -------------------------------------------------------------------------------- /lib/model/roi_pooling/modules/roi_pool.py: -------------------------------------------------------------------------------- 1 | from torch.nn.modules.module import Module 2 | from ..functions.roi_pool import RoIPoolFunction 3 | 4 | 5 | class _RoIPooling(Module): 6 | def __init__(self, pooled_height, pooled_width, spatial_scale): 7 | super(_RoIPooling, self).__init__() 8 | 9 | self.pooled_width = int(pooled_width) 10 | self.pooled_height = int(pooled_height) 11 | self.spatial_scale = float(spatial_scale) 12 | 13 | def forward(self, features, rois): 14 | return RoIPoolFunction(self.pooled_height, self.pooled_width, self.spatial_scale)(features, rois) 15 | -------------------------------------------------------------------------------- /lib/model/csrc/cpu/vision.h: -------------------------------------------------------------------------------- 1 | // Copyright (c) Facebook, Inc. and its affiliates. All Rights Reserved. 2 | #pragma once 3 | #include 4 | 5 | 6 | at::Tensor ROIAlign_forward_cpu(const at::Tensor& input, 7 | const at::Tensor& rois, 8 | const float spatial_scale, 9 | const int pooled_height, 10 | const int pooled_width, 11 | const int sampling_ratio); 12 | 13 | 14 | at::Tensor nms_cpu(const at::Tensor& dets, 15 | const at::Tensor& scores, 16 | const float threshold); 17 | -------------------------------------------------------------------------------- /lib/model/roi_crop/src/roi_crop.h: -------------------------------------------------------------------------------- 1 | int BilinearSamplerBHWD_updateOutput(THFloatTensor *inputImages, THFloatTensor *grids, THFloatTensor *output); 2 | 3 | int BilinearSamplerBHWD_updateGradInput(THFloatTensor *inputImages, THFloatTensor *grids, THFloatTensor *gradInputImages, 4 | THFloatTensor *gradGrids, THFloatTensor *gradOutput); 5 | 6 | 7 | 8 | int BilinearSamplerBCHW_updateOutput(THFloatTensor *inputImages, THFloatTensor *grids, THFloatTensor *output); 9 | 10 | int BilinearSamplerBCHW_updateGradInput(THFloatTensor *inputImages, THFloatTensor *grids, THFloatTensor *gradInputImages, 11 | THFloatTensor *gradGrids, THFloatTensor *gradOutput); 12 | -------------------------------------------------------------------------------- /lib/model/csrc/nms.h: -------------------------------------------------------------------------------- 1 | // Copyright (c) Facebook, Inc. and its affiliates. All Rights Reserved. 2 | #pragma once 3 | #include "cpu/vision.h" 4 | 5 | #ifdef WITH_CUDA 6 | #include "cuda/vision.h" 7 | #endif 8 | 9 | 10 | at::Tensor nms(const at::Tensor& dets, 11 | const at::Tensor& scores, 12 | const float threshold) { 13 | 14 | if (dets.type().is_cuda()) { 15 | #ifdef WITH_CUDA 16 | // TODO raise error if not compiled with CUDA 17 | if (dets.numel() == 0) 18 | return at::empty({0}, dets.options().dtype(at::kLong).device(at::kCPU)); 19 | auto b = at::cat({dets, scores.unsqueeze(1)}, 1); 20 | return nms_cuda(b, threshold); 21 | #else 22 | AT_ERROR("Not compiled with GPU support"); 23 | #endif 24 | } 25 | 26 | at::Tensor result = nms_cpu(dets, scores, threshold); 27 | return result; 28 | } 29 | -------------------------------------------------------------------------------- /lib/model/nms/nms_wrapper.py: -------------------------------------------------------------------------------- 1 | # -------------------------------------------------------- 2 | # Fast R-CNN 3 | # Copyright (c) 2015 Microsoft 4 | # Licensed under The MIT License [see LICENSE for details] 5 | # Written by Ross Girshick 6 | # -------------------------------------------------------- 7 | import torch 8 | from model.utils.config import cfg 9 | if torch.cuda.is_available(): 10 | from model.nms.nms_gpu import nms_gpu 11 | from model.nms.nms_cpu import nms_cpu 12 | 13 | def nms(dets, thresh, force_cpu=False): 14 | """Dispatch to either CPU or GPU NMS implementations.""" 15 | if dets.shape[0] == 0: 16 | return [] 17 | # ---numpy version--- 18 | # original: return gpu_nms(dets, thresh, device_id=cfg.GPU_ID) 19 | # ---pytorch version--- 20 | 21 | return nms_gpu(dets, thresh) if force_cpu == False else nms_cpu(dets, thresh) 22 | -------------------------------------------------------------------------------- /lib/model/roi_pooling/src/roi_pooling_kernel.h: -------------------------------------------------------------------------------- 1 | #ifndef _ROI_POOLING_KERNEL 2 | #define _ROI_POOLING_KERNEL 3 | 4 | #ifdef __cplusplus 5 | extern "C" { 6 | #endif 7 | 8 | int ROIPoolForwardLaucher( 9 | const float* bottom_data, const float spatial_scale, const int num_rois, const int height, 10 | const int width, const int channels, const int pooled_height, 11 | const int pooled_width, const float* bottom_rois, 12 | float* top_data, int* argmax_data, cudaStream_t stream); 13 | 14 | 15 | int ROIPoolBackwardLaucher(const float* top_diff, const float spatial_scale, const int batch_size, const int num_rois, 16 | const int height, const int width, const int channels, const int pooled_height, 17 | const int pooled_width, const float* bottom_rois, 18 | float* bottom_diff, const int* argmax_data, cudaStream_t stream); 19 | 20 | #ifdef __cplusplus 21 | } 22 | #endif 23 | 24 | #endif 25 | 26 | -------------------------------------------------------------------------------- /lib/model/nms/build.py: -------------------------------------------------------------------------------- 1 | from __future__ import print_function 2 | import os 3 | import torch 4 | from torch.utils.ffi import create_extension 5 | 6 | #this_file = os.path.dirname(__file__) 7 | 8 | sources = [] 9 | headers = [] 10 | defines = [] 11 | with_cuda = False 12 | 13 | if torch.cuda.is_available(): 14 | print('Including CUDA code.') 15 | sources += ['src/nms_cuda.c'] 16 | headers += ['src/nms_cuda.h'] 17 | defines += [('WITH_CUDA', None)] 18 | with_cuda = True 19 | 20 | this_file = os.path.dirname(os.path.realpath(__file__)) 21 | print(this_file) 22 | extra_objects = ['src/nms_cuda_kernel.cu.o'] 23 | extra_objects = [os.path.join(this_file, fname) for fname in extra_objects] 24 | print(extra_objects) 25 | 26 | ffi = create_extension( 27 | '_ext.nms', 28 | headers=headers, 29 | sources=sources, 30 | define_macros=defines, 31 | relative_to=__file__, 32 | with_cuda=with_cuda, 33 | extra_objects=extra_objects 34 | ) 35 | 36 | if __name__ == '__main__': 37 | ffi.build() 38 | -------------------------------------------------------------------------------- /lib/model/nms/nms_cpu.py: -------------------------------------------------------------------------------- 1 | from __future__ import absolute_import 2 | 3 | import numpy as np 4 | import torch 5 | 6 | def nms_cpu(dets, thresh): 7 | dets = dets.numpy() 8 | x1 = dets[:, 0] 9 | y1 = dets[:, 1] 10 | x2 = dets[:, 2] 11 | y2 = dets[:, 3] 12 | scores = dets[:, 4] 13 | 14 | areas = (x2 - x1 + 1) * (y2 - y1 + 1) 15 | order = scores.argsort()[::-1] 16 | 17 | keep = [] 18 | while order.size > 0: 19 | i = order.item(0) 20 | keep.append(i) 21 | xx1 = np.maximum(x1[i], x1[order[1:]]) 22 | yy1 = np.maximum(y1[i], y1[order[1:]]) 23 | xx2 = np.maximum(x2[i], x2[order[1:]]) 24 | yy2 = np.maximum(y2[i], y2[order[1:]]) 25 | 26 | w = np.maximum(0.0, xx2 - xx1 + 1) 27 | h = np.maximum(0.0, yy2 - yy1 + 1) 28 | inter = w * h 29 | ovr = inter / (areas[i] + areas[order[1:]] - inter) 30 | 31 | inds = np.where(ovr <= thresh)[0] 32 | order = order[inds + 1] 33 | 34 | return torch.IntTensor(keep) 35 | 36 | 37 | -------------------------------------------------------------------------------- /lib/model/roi_pooling/build.py: -------------------------------------------------------------------------------- 1 | from __future__ import print_function 2 | import os 3 | import torch 4 | from torch.utils.ffi import create_extension 5 | 6 | 7 | sources = ['src/roi_pooling.c'] 8 | headers = ['src/roi_pooling.h'] 9 | extra_objects = [] 10 | defines = [] 11 | with_cuda = False 12 | 13 | this_file = os.path.dirname(os.path.realpath(__file__)) 14 | print(this_file) 15 | 16 | if torch.cuda.is_available(): 17 | print('Including CUDA code.') 18 | sources += ['src/roi_pooling_cuda.c'] 19 | headers += ['src/roi_pooling_cuda.h'] 20 | defines += [('WITH_CUDA', None)] 21 | with_cuda = True 22 | extra_objects = ['src/roi_pooling.cu.o'] 23 | extra_objects = [os.path.join(this_file, fname) for fname in extra_objects] 24 | 25 | ffi = create_extension( 26 | '_ext.roi_pooling', 27 | headers=headers, 28 | sources=sources, 29 | define_macros=defines, 30 | relative_to=__file__, 31 | with_cuda=with_cuda, 32 | extra_objects=extra_objects 33 | ) 34 | 35 | if __name__ == '__main__': 36 | ffi.build() 37 | -------------------------------------------------------------------------------- /lib/model/roi_crop/build.py: -------------------------------------------------------------------------------- 1 | from __future__ import print_function 2 | import os 3 | import torch 4 | from torch.utils.ffi import create_extension 5 | 6 | #this_file = os.path.dirname(__file__) 7 | 8 | sources = ['src/roi_crop.c'] 9 | headers = ['src/roi_crop.h'] 10 | defines = [] 11 | with_cuda = False 12 | 13 | if torch.cuda.is_available(): 14 | print('Including CUDA code.') 15 | sources += ['src/roi_crop_cuda.c'] 16 | headers += ['src/roi_crop_cuda.h'] 17 | defines += [('WITH_CUDA', None)] 18 | with_cuda = True 19 | 20 | this_file = os.path.dirname(os.path.realpath(__file__)) 21 | print(this_file) 22 | extra_objects = ['src/roi_crop_cuda_kernel.cu.o'] 23 | extra_objects = [os.path.join(this_file, fname) for fname in extra_objects] 24 | 25 | ffi = create_extension( 26 | '_ext.roi_crop', 27 | headers=headers, 28 | sources=sources, 29 | define_macros=defines, 30 | relative_to=__file__, 31 | with_cuda=with_cuda, 32 | extra_objects=extra_objects 33 | ) 34 | 35 | if __name__ == '__main__': 36 | ffi.build() 37 | -------------------------------------------------------------------------------- /lib/model/roi_align/build.py: -------------------------------------------------------------------------------- 1 | from __future__ import print_function 2 | import os 3 | import torch 4 | from torch.utils.ffi import create_extension 5 | 6 | sources = ['src/roi_align.c'] 7 | headers = ['src/roi_align.h'] 8 | extra_objects = [] 9 | #sources = [] 10 | #headers = [] 11 | defines = [] 12 | with_cuda = False 13 | 14 | this_file = os.path.dirname(os.path.realpath(__file__)) 15 | print(this_file) 16 | 17 | if torch.cuda.is_available(): 18 | print('Including CUDA code.') 19 | sources += ['src/roi_align_cuda.c'] 20 | headers += ['src/roi_align_cuda.h'] 21 | defines += [('WITH_CUDA', None)] 22 | with_cuda = True 23 | 24 | extra_objects = ['src/roi_align_kernel.cu.o'] 25 | extra_objects = [os.path.join(this_file, fname) for fname in extra_objects] 26 | 27 | ffi = create_extension( 28 | '_ext.roi_align', 29 | headers=headers, 30 | sources=sources, 31 | define_macros=defines, 32 | relative_to=__file__, 33 | with_cuda=with_cuda, 34 | extra_objects=extra_objects 35 | ) 36 | 37 | if __name__ == '__main__': 38 | ffi.build() 39 | -------------------------------------------------------------------------------- /lib/model/roi_crop/functions/roi_crop.py: -------------------------------------------------------------------------------- 1 | # functions/add.py 2 | import torch 3 | from torch.autograd import Function 4 | from .._ext import roi_crop 5 | import pdb 6 | 7 | class RoICropFunction(Function): 8 | def forward(self, input1, input2): 9 | self.input1 = input1.clone() 10 | self.input2 = input2.clone() 11 | output = input2.new(input2.size()[0], input1.size()[1], input2.size()[1], input2.size()[2]).zero_() 12 | assert output.get_device() == input1.get_device(), "output and input1 must on the same device" 13 | assert output.get_device() == input2.get_device(), "output and input2 must on the same device" 14 | roi_crop.BilinearSamplerBHWD_updateOutput_cuda(input1, input2, output) 15 | return output 16 | 17 | def backward(self, grad_output): 18 | grad_input1 = self.input1.new(self.input1.size()).zero_() 19 | grad_input2 = self.input2.new(self.input2.size()).zero_() 20 | roi_crop.BilinearSamplerBHWD_updateGradInput_cuda(self.input1, self.input2, grad_input1, grad_input2, grad_output) 21 | return grad_input1, grad_input2 22 | -------------------------------------------------------------------------------- /LICENSE: -------------------------------------------------------------------------------- 1 | MIT License 2 | 3 | Copyright (c) 2020 Dandan Shan 4 | 5 | Permission is hereby granted, free of charge, to any person obtaining a copy 6 | of this software and associated documentation files (the "Software"), to deal 7 | in the Software without restriction, including without limitation the rights 8 | to use, copy, modify, merge, publish, distribute, sublicense, and/or sell 9 | copies of the Software, and to permit persons to whom the Software is 10 | furnished to do so, subject to the following conditions: 11 | 12 | The above copyright notice and this permission notice shall be included in all 13 | copies or substantial portions of the Software. 14 | 15 | THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR 16 | IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, 17 | FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE 18 | AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER 19 | LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, 20 | OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE 21 | SOFTWARE. 22 | -------------------------------------------------------------------------------- /lib/model/roi_align/src/roi_align_kernel.h: -------------------------------------------------------------------------------- 1 | #ifndef _ROI_ALIGN_KERNEL 2 | #define _ROI_ALIGN_KERNEL 3 | 4 | #ifdef __cplusplus 5 | extern "C" { 6 | #endif 7 | 8 | __global__ void ROIAlignForward(const int nthreads, const float* bottom_data, 9 | const float spatial_scale, const int height, const int width, 10 | const int channels, const int aligned_height, const int aligned_width, 11 | const float* bottom_rois, float* top_data); 12 | 13 | int ROIAlignForwardLaucher( 14 | const float* bottom_data, const float spatial_scale, const int num_rois, const int height, 15 | const int width, const int channels, const int aligned_height, 16 | const int aligned_width, const float* bottom_rois, 17 | float* top_data, cudaStream_t stream); 18 | 19 | __global__ void ROIAlignBackward(const int nthreads, const float* top_diff, 20 | const float spatial_scale, const int height, const int width, 21 | const int channels, const int aligned_height, const int aligned_width, 22 | float* bottom_diff, const float* bottom_rois); 23 | 24 | int ROIAlignBackwardLaucher(const float* top_diff, const float spatial_scale, const int batch_size, const int num_rois, 25 | const int height, const int width, const int channels, const int aligned_height, 26 | const int aligned_width, const float* bottom_rois, 27 | float* bottom_diff, cudaStream_t stream); 28 | 29 | #ifdef __cplusplus 30 | } 31 | #endif 32 | 33 | #endif 34 | 35 | -------------------------------------------------------------------------------- /lib/datasets/VOCdevkit-matlab-wrapper/voc_eval.m: -------------------------------------------------------------------------------- 1 | function res = voc_eval(path, comp_id, test_set, output_dir) 2 | 3 | VOCopts = get_voc_opts(path); 4 | VOCopts.testset = test_set; 5 | 6 | for i = 1:length(VOCopts.classes) 7 | cls = VOCopts.classes{i}; 8 | res(i) = voc_eval_cls(cls, VOCopts, comp_id, output_dir); 9 | end 10 | 11 | fprintf('\n~~~~~~~~~~~~~~~~~~~~\n'); 12 | fprintf('Results:\n'); 13 | aps = [res(:).ap]'; 14 | fprintf('%.1f\n', aps * 100); 15 | fprintf('%.1f\n', mean(aps) * 100); 16 | fprintf('~~~~~~~~~~~~~~~~~~~~\n'); 17 | 18 | function res = voc_eval_cls(cls, VOCopts, comp_id, output_dir) 19 | 20 | test_set = VOCopts.testset; 21 | year = VOCopts.dataset(4:end); 22 | 23 | addpath(fullfile(VOCopts.datadir, 'VOCcode')); 24 | 25 | res_fn = sprintf(VOCopts.detrespath, comp_id, cls); 26 | 27 | recall = []; 28 | prec = []; 29 | ap = 0; 30 | ap_auc = 0; 31 | 32 | do_eval = (str2num(year) <= 2007) | ~strcmp(test_set, 'test'); 33 | if do_eval 34 | % Bug in VOCevaldet requires that tic has been called first 35 | tic; 36 | [recall, prec, ap] = VOCevaldet(VOCopts, comp_id, cls, true); 37 | ap_auc = xVOCap(recall, prec); 38 | 39 | % force plot limits 40 | ylim([0 1]); 41 | xlim([0 1]); 42 | 43 | print(gcf, '-djpeg', '-r0', ... 44 | [output_dir '/' cls '_pr.jpg']); 45 | end 46 | fprintf('!!! %s : %.4f %.4f\n', cls, ap, ap_auc); 47 | 48 | res.recall = recall; 49 | res.prec = prec; 50 | res.ap = ap; 51 | res.ap_auc = ap_auc; 52 | 53 | save([output_dir '/' cls '_pr.mat'], ... 54 | 'res', 'recall', 'prec', 'ap', 'ap_auc'); 55 | 56 | rmpath(fullfile(VOCopts.datadir, 'VOCcode')); 57 | -------------------------------------------------------------------------------- /lib/datasets/ds_utils.py: -------------------------------------------------------------------------------- 1 | # -------------------------------------------------------- 2 | # Fast/er R-CNN 3 | # Licensed under The MIT License [see LICENSE for details] 4 | # Written by Ross Girshick 5 | # -------------------------------------------------------- 6 | from __future__ import absolute_import 7 | from __future__ import division 8 | from __future__ import print_function 9 | 10 | import numpy as np 11 | 12 | 13 | def unique_boxes(boxes, scale=1.0): 14 | """Return indices of unique boxes.""" 15 | v = np.array([1, 1e3, 1e6, 1e9]) 16 | hashes = np.round(boxes * scale).dot(v) 17 | _, index = np.unique(hashes, return_index=True) 18 | return np.sort(index) 19 | 20 | 21 | def xywh_to_xyxy(boxes): 22 | """Convert [x y w h] box format to [x1 y1 x2 y2] format.""" 23 | return np.hstack((boxes[:, 0:2], boxes[:, 0:2] + boxes[:, 2:4] - 1)) 24 | 25 | 26 | def xyxy_to_xywh(boxes): 27 | """Convert [x1 y1 x2 y2] box format to [x y w h] format.""" 28 | return np.hstack((boxes[:, 0:2], boxes[:, 2:4] - boxes[:, 0:2] + 1)) 29 | 30 | 31 | def validate_boxes(boxes, width=0, height=0): 32 | """Check that a set of boxes are valid.""" 33 | x1 = boxes[:, 0] 34 | y1 = boxes[:, 1] 35 | x2 = boxes[:, 2] 36 | y2 = boxes[:, 3] 37 | assert (x1 >= 0).all() 38 | assert (y1 >= 0).all() 39 | assert (x2 >= x1).all() 40 | assert (y2 >= y1).all() 41 | assert (x2 < width).all() 42 | assert (y2 < height).all() 43 | 44 | 45 | def filter_small_boxes(boxes, min_size): 46 | w = boxes[:, 2] - boxes[:, 0] 47 | h = boxes[:, 3] - boxes[:, 1] 48 | keep = np.where((w >= min_size) & (h > min_size))[0] 49 | return keep 50 | -------------------------------------------------------------------------------- /lib/pycocotools/license.txt: -------------------------------------------------------------------------------- 1 | Copyright (c) 2014, Piotr Dollar and Tsung-Yi Lin 2 | All rights reserved. 3 | 4 | Redistribution and use in source and binary forms, with or without 5 | modification, are permitted provided that the following conditions are met: 6 | 7 | 1. Redistributions of source code must retain the above copyright notice, this 8 | list of conditions and the following disclaimer. 9 | 2. Redistributions in binary form must reproduce the above copyright notice, 10 | this list of conditions and the following disclaimer in the documentation 11 | and/or other materials provided with the distribution. 12 | 13 | THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND 14 | ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED 15 | WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE 16 | DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR 17 | ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES 18 | (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; 19 | LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND 20 | ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT 21 | (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS 22 | SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. 23 | 24 | The views and conclusions contained in the software and documentation are those 25 | of the authors and should not be interpreted as representing official policies, 26 | either expressed or implied, of the FreeBSD Project. 27 | -------------------------------------------------------------------------------- /lib/datasets/tools/mcg_munge.py: -------------------------------------------------------------------------------- 1 | from __future__ import print_function 2 | import os 3 | import sys 4 | 5 | """Hacky tool to convert file system layout of MCG boxes downloaded from 6 | http://www.eecs.berkeley.edu/Research/Projects/CS/vision/grouping/mcg/ 7 | so that it's consistent with those computed by Jan Hosang (see: 8 | http://www.mpi-inf.mpg.de/departments/computer-vision-and-multimodal- 9 | computing/research/object-recognition-and-scene-understanding/how- 10 | good-are-detection-proposals-really/) 11 | 12 | NB: Boxes from the MCG website are in (y1, x1, y2, x2) order. 13 | Boxes from Hosang et al. are in (x1, y1, x2, y2) order. 14 | """ 15 | 16 | def munge(src_dir): 17 | # stored as: ./MCG-COCO-val2014-boxes/COCO_val2014_000000193401.mat 18 | # want: ./MCG/mat/COCO_val2014_0/COCO_val2014_000000141/COCO_val2014_000000141334.mat 19 | 20 | files = os.listdir(src_dir) 21 | for fn in files: 22 | base, ext = os.path.splitext(fn) 23 | # first 14 chars / first 22 chars / all chars + .mat 24 | # COCO_val2014_0/COCO_val2014_000000447/COCO_val2014_000000447991.mat 25 | first = base[:14] 26 | second = base[:22] 27 | dst_dir = os.path.join('MCG', 'mat', first, second) 28 | if not os.path.exists(dst_dir): 29 | os.makedirs(dst_dir) 30 | src = os.path.join(src_dir, fn) 31 | dst = os.path.join(dst_dir, fn) 32 | print('MV: {} -> {}'.format(src, dst)) 33 | os.rename(src, dst) 34 | 35 | if __name__ == '__main__': 36 | # src_dir should look something like: 37 | # src_dir = 'MCG-COCO-val2014-boxes' 38 | src_dir = sys.argv[1] 39 | munge(src_dir) 40 | -------------------------------------------------------------------------------- /lib/model/roi_crop/functions/crop_resize.py: -------------------------------------------------------------------------------- 1 | # functions/add.py 2 | import torch 3 | from torch.autograd import Function 4 | from .._ext import roi_crop 5 | from cffi import FFI 6 | ffi = FFI() 7 | 8 | class RoICropFunction(Function): 9 | def forward(self, input1, input2): 10 | self.input1 = input1 11 | self.input2 = input2 12 | self.device_c = ffi.new("int *") 13 | output = torch.zeros(input2.size()[0], input1.size()[1], input2.size()[1], input2.size()[2]) 14 | #print('decice %d' % torch.cuda.current_device()) 15 | if input1.is_cuda: 16 | self.device = torch.cuda.current_device() 17 | else: 18 | self.device = -1 19 | self.device_c[0] = self.device 20 | if not input1.is_cuda: 21 | roi_crop.BilinearSamplerBHWD_updateOutput(input1, input2, output) 22 | else: 23 | output = output.cuda(self.device) 24 | roi_crop.BilinearSamplerBHWD_updateOutput_cuda(input1, input2, output) 25 | return output 26 | 27 | def backward(self, grad_output): 28 | grad_input1 = torch.zeros(self.input1.size()) 29 | grad_input2 = torch.zeros(self.input2.size()) 30 | #print('backward decice %d' % self.device) 31 | if not grad_output.is_cuda: 32 | roi_crop.BilinearSamplerBHWD_updateGradInput(self.input1, self.input2, grad_input1, grad_input2, grad_output) 33 | else: 34 | grad_input1 = grad_input1.cuda(self.device) 35 | grad_input2 = grad_input2.cuda(self.device) 36 | roi_crop.BilinearSamplerBHWD_updateGradInput_cuda(self.input1, self.input2, grad_input1, grad_input2, grad_output) 37 | return grad_input1, grad_input2 38 | -------------------------------------------------------------------------------- /lib/model/csrc/ROIPool.h: -------------------------------------------------------------------------------- 1 | // Copyright (c) Facebook, Inc. and its affiliates. All Rights Reserved. 2 | #pragma once 3 | 4 | #include "cpu/vision.h" 5 | 6 | #ifdef WITH_CUDA 7 | #include "cuda/vision.h" 8 | #endif 9 | 10 | 11 | std::tuple ROIPool_forward(const at::Tensor& input, 12 | const at::Tensor& rois, 13 | const float spatial_scale, 14 | const int pooled_height, 15 | const int pooled_width) { 16 | if (input.type().is_cuda()) { 17 | #ifdef WITH_CUDA 18 | return ROIPool_forward_cuda(input, rois, spatial_scale, pooled_height, pooled_width); 19 | #else 20 | AT_ERROR("Not compiled with GPU support"); 21 | #endif 22 | } 23 | AT_ERROR("Not implemented on the CPU"); 24 | } 25 | 26 | at::Tensor ROIPool_backward(const at::Tensor& grad, 27 | const at::Tensor& input, 28 | const at::Tensor& rois, 29 | const at::Tensor& argmax, 30 | const float spatial_scale, 31 | const int pooled_height, 32 | const int pooled_width, 33 | const int batch_size, 34 | const int channels, 35 | const int height, 36 | const int width) { 37 | if (grad.type().is_cuda()) { 38 | #ifdef WITH_CUDA 39 | return ROIPool_backward_cuda(grad, input, rois, argmax, spatial_scale, pooled_height, pooled_width, batch_size, channels, height, width); 40 | #else 41 | AT_ERROR("Not compiled with GPU support"); 42 | #endif 43 | } 44 | AT_ERROR("Not implemented on the CPU"); 45 | } 46 | 47 | 48 | 49 | -------------------------------------------------------------------------------- /lib/model/csrc/ROIAlign.h: -------------------------------------------------------------------------------- 1 | // Copyright (c) Facebook, Inc. and its affiliates. All Rights Reserved. 2 | #pragma once 3 | 4 | #include "cpu/vision.h" 5 | 6 | #ifdef WITH_CUDA 7 | #include "cuda/vision.h" 8 | #endif 9 | 10 | // Interface for Python 11 | at::Tensor ROIAlign_forward(const at::Tensor& input, 12 | const at::Tensor& rois, 13 | const float spatial_scale, 14 | const int pooled_height, 15 | const int pooled_width, 16 | const int sampling_ratio) { 17 | if (input.type().is_cuda()) { 18 | #ifdef WITH_CUDA 19 | return ROIAlign_forward_cuda(input, rois, spatial_scale, pooled_height, pooled_width, sampling_ratio); 20 | #else 21 | AT_ERROR("Not compiled with GPU support"); 22 | #endif 23 | } 24 | return ROIAlign_forward_cpu(input, rois, spatial_scale, pooled_height, pooled_width, sampling_ratio); 25 | } 26 | 27 | at::Tensor ROIAlign_backward(const at::Tensor& grad, 28 | const at::Tensor& rois, 29 | const float spatial_scale, 30 | const int pooled_height, 31 | const int pooled_width, 32 | const int batch_size, 33 | const int channels, 34 | const int height, 35 | const int width, 36 | const int sampling_ratio) { 37 | if (grad.type().is_cuda()) { 38 | #ifdef WITH_CUDA 39 | return ROIAlign_backward_cuda(grad, rois, spatial_scale, pooled_height, pooled_width, batch_size, channels, height, width, sampling_ratio); 40 | #else 41 | AT_ERROR("Not compiled with GPU support"); 42 | #endif 43 | } 44 | AT_ERROR("Not implemented on the CPU"); 45 | } 46 | 47 | -------------------------------------------------------------------------------- /lib/model/roi_align/modules/roi_align.py: -------------------------------------------------------------------------------- 1 | from torch.nn.modules.module import Module 2 | from torch.nn.functional import avg_pool2d, max_pool2d 3 | from ..functions.roi_align import RoIAlignFunction 4 | 5 | 6 | class RoIAlign(Module): 7 | def __init__(self, aligned_height, aligned_width, spatial_scale): 8 | super(RoIAlign, self).__init__() 9 | 10 | self.aligned_width = int(aligned_width) 11 | self.aligned_height = int(aligned_height) 12 | self.spatial_scale = float(spatial_scale) 13 | 14 | def forward(self, features, rois): 15 | return RoIAlignFunction(self.aligned_height, self.aligned_width, 16 | self.spatial_scale)(features, rois) 17 | 18 | class RoIAlignAvg(Module): 19 | def __init__(self, aligned_height, aligned_width, spatial_scale): 20 | super(RoIAlignAvg, self).__init__() 21 | 22 | self.aligned_width = int(aligned_width) 23 | self.aligned_height = int(aligned_height) 24 | self.spatial_scale = float(spatial_scale) 25 | 26 | def forward(self, features, rois): 27 | x = RoIAlignFunction(self.aligned_height+1, self.aligned_width+1, 28 | self.spatial_scale)(features, rois) 29 | return avg_pool2d(x, kernel_size=2, stride=1) 30 | 31 | class RoIAlignMax(Module): 32 | def __init__(self, aligned_height, aligned_width, spatial_scale): 33 | super(RoIAlignMax, self).__init__() 34 | 35 | self.aligned_width = int(aligned_width) 36 | self.aligned_height = int(aligned_height) 37 | self.spatial_scale = float(spatial_scale) 38 | 39 | def forward(self, features, rois): 40 | x = RoIAlignFunction(self.aligned_height+1, self.aligned_width+1, 41 | self.spatial_scale)(features, rois) 42 | return max_pool2d(x, kernel_size=2, stride=1) 43 | -------------------------------------------------------------------------------- /lib/model/utils/blob.py: -------------------------------------------------------------------------------- 1 | # -------------------------------------------------------- 2 | # Fast R-CNN 3 | # Copyright (c) 2015 Microsoft 4 | # Licensed under The MIT License [see LICENSE for details] 5 | # Written by Ross Girshick 6 | # -------------------------------------------------------- 7 | 8 | """Blob helper functions.""" 9 | 10 | import numpy as np 11 | # from scipy.misc import imread, imresize 12 | import cv2 13 | 14 | try: 15 | xrange # Python 2 16 | except NameError: 17 | xrange = range # Python 3 18 | 19 | 20 | def im_list_to_blob(ims): 21 | """Convert a list of images into a network input. 22 | 23 | Assumes images are already prepared (means subtracted, BGR order, ...). 24 | """ 25 | max_shape = np.array([im.shape for im in ims]).max(axis=0) 26 | num_images = len(ims) 27 | blob = np.zeros((num_images, max_shape[0], max_shape[1], 3), 28 | dtype=np.float32) 29 | for i in xrange(num_images): 30 | im = ims[i] 31 | blob[i, 0:im.shape[0], 0:im.shape[1], :] = im 32 | 33 | return blob 34 | 35 | def prep_im_for_blob(im, pixel_means, target_size, max_size): 36 | """Mean subtract and scale an image for use in a blob.""" 37 | 38 | im = im.astype(np.float32, copy=False) 39 | im -= pixel_means 40 | # im = im[:, :, ::-1] 41 | im_shape = im.shape 42 | im_size_min = np.min(im_shape[0:2]) 43 | im_size_max = np.max(im_shape[0:2]) 44 | im_scale = float(target_size) / float(im_size_min) 45 | # Prevent the biggest axis from being more than MAX_SIZE 46 | # if np.round(im_scale * im_size_max) > max_size: 47 | # im_scale = float(max_size) / float(im_size_max) 48 | # im = imresize(im, im_scale) 49 | im = cv2.resize(im, None, None, fx=im_scale, fy=im_scale, 50 | interpolation=cv2.INTER_LINEAR) 51 | 52 | return im, im_scale 53 | -------------------------------------------------------------------------------- /lib/model/roi_pooling/functions/roi_pool.py: -------------------------------------------------------------------------------- 1 | import torch 2 | from torch.autograd import Function 3 | from .._ext import roi_pooling 4 | import pdb 5 | 6 | class RoIPoolFunction(Function): 7 | def __init__(ctx, pooled_height, pooled_width, spatial_scale): 8 | ctx.pooled_width = pooled_width 9 | ctx.pooled_height = pooled_height 10 | ctx.spatial_scale = spatial_scale 11 | ctx.feature_size = None 12 | 13 | def forward(ctx, features, rois): 14 | ctx.feature_size = features.size() 15 | batch_size, num_channels, data_height, data_width = ctx.feature_size 16 | num_rois = rois.size(0) 17 | output = features.new(num_rois, num_channels, ctx.pooled_height, ctx.pooled_width).zero_() 18 | ctx.argmax = features.new(num_rois, num_channels, ctx.pooled_height, ctx.pooled_width).zero_().int() 19 | ctx.rois = rois 20 | if not features.is_cuda: 21 | _features = features.permute(0, 2, 3, 1) 22 | roi_pooling.roi_pooling_forward(ctx.pooled_height, ctx.pooled_width, ctx.spatial_scale, 23 | _features, rois, output) 24 | else: 25 | roi_pooling.roi_pooling_forward_cuda(ctx.pooled_height, ctx.pooled_width, ctx.spatial_scale, 26 | features, rois, output, ctx.argmax) 27 | 28 | return output 29 | 30 | def backward(ctx, grad_output): 31 | assert(ctx.feature_size is not None and grad_output.is_cuda) 32 | batch_size, num_channels, data_height, data_width = ctx.feature_size 33 | grad_input = grad_output.new(batch_size, num_channels, data_height, data_width).zero_() 34 | 35 | roi_pooling.roi_pooling_backward_cuda(ctx.pooled_height, ctx.pooled_width, ctx.spatial_scale, 36 | grad_output, ctx.rois, grad_input, ctx.argmax) 37 | 38 | return grad_input, None 39 | -------------------------------------------------------------------------------- /lib/model/roi_layers/roi_pool.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) Facebook, Inc. and its affiliates. All Rights Reserved. 2 | import torch 3 | from torch import nn 4 | from torch.autograd import Function 5 | from torch.autograd.function import once_differentiable 6 | from torch.nn.modules.utils import _pair 7 | 8 | from model import _C 9 | 10 | 11 | class _ROIPool(Function): 12 | @staticmethod 13 | def forward(ctx, input, roi, output_size, spatial_scale): 14 | ctx.output_size = _pair(output_size) 15 | ctx.spatial_scale = spatial_scale 16 | ctx.input_shape = input.size() 17 | output, argmax = _C.roi_pool_forward( 18 | input, roi, spatial_scale, output_size[0], output_size[1] 19 | ) 20 | ctx.save_for_backward(input, roi, argmax) 21 | return output 22 | 23 | @staticmethod 24 | @once_differentiable 25 | def backward(ctx, grad_output): 26 | input, rois, argmax = ctx.saved_tensors 27 | output_size = ctx.output_size 28 | spatial_scale = ctx.spatial_scale 29 | bs, ch, h, w = ctx.input_shape 30 | grad_input = _C.roi_pool_backward( 31 | grad_output, 32 | input, 33 | rois, 34 | argmax, 35 | spatial_scale, 36 | output_size[0], 37 | output_size[1], 38 | bs, 39 | ch, 40 | h, 41 | w, 42 | ) 43 | return grad_input, None, None, None 44 | 45 | 46 | roi_pool = _ROIPool.apply 47 | 48 | 49 | class ROIPool(nn.Module): 50 | def __init__(self, output_size, spatial_scale): 51 | super(ROIPool, self).__init__() 52 | self.output_size = output_size 53 | self.spatial_scale = spatial_scale 54 | 55 | def forward(self, input, rois): 56 | return roi_pool(input, rois, self.output_size, self.spatial_scale) 57 | 58 | def __repr__(self): 59 | tmpstr = self.__class__.__name__ + "(" 60 | tmpstr += "output_size=" + str(self.output_size) 61 | tmpstr += ", spatial_scale=" + str(self.spatial_scale) 62 | tmpstr += ")" 63 | return tmpstr 64 | -------------------------------------------------------------------------------- /lib/pycocotools/maskApi.h: -------------------------------------------------------------------------------- 1 | /************************************************************************** 2 | * Microsoft COCO Toolbox. version 2.0 3 | * Data, paper, and tutorials available at: http://mscoco.org/ 4 | * Code written by Piotr Dollar and Tsung-Yi Lin, 2015. 5 | * Licensed under the Simplified BSD License [see coco/license.txt] 6 | **************************************************************************/ 7 | #pragma once 8 | #include 9 | 10 | typedef unsigned int uint; 11 | typedef unsigned long siz; 12 | typedef unsigned char byte; 13 | typedef double* BB; 14 | typedef struct { siz h, w, m; uint *cnts; } RLE; 15 | 16 | // Initialize/destroy RLE. 17 | void rleInit( RLE *R, siz h, siz w, siz m, uint *cnts ); 18 | void rleFree( RLE *R ); 19 | 20 | // Initialize/destroy RLE array. 21 | void rlesInit( RLE **R, siz n ); 22 | void rlesFree( RLE **R, siz n ); 23 | 24 | // Encode binary masks using RLE. 25 | void rleEncode( RLE *R, const byte *mask, siz h, siz w, siz n ); 26 | 27 | // Decode binary masks encoded via RLE. 28 | void rleDecode( const RLE *R, byte *mask, siz n ); 29 | 30 | // Compute union or intersection of encoded masks. 31 | void rleMerge( const RLE *R, RLE *M, siz n, bool intersect ); 32 | 33 | // Compute area of encoded masks. 34 | void rleArea( const RLE *R, siz n, uint *a ); 35 | 36 | // Compute intersection over union between masks. 37 | void rleIou( RLE *dt, RLE *gt, siz m, siz n, byte *iscrowd, double *o ); 38 | 39 | // Compute intersection over union between bounding boxes. 40 | void bbIou( BB dt, BB gt, siz m, siz n, byte *iscrowd, double *o ); 41 | 42 | // Get bounding boxes surrounding encoded masks. 43 | void rleToBbox( const RLE *R, BB bb, siz n ); 44 | 45 | // Convert bounding boxes to encoded masks. 46 | void rleFrBbox( RLE *R, const BB bb, siz h, siz w, siz n ); 47 | 48 | // Convert polygon to encoded mask. 49 | void rleFrPoly( RLE *R, const double *xy, siz k, siz h, siz w ); 50 | 51 | // Get compressed string representation of encoded mask. 52 | char* rleToString( const RLE *R ); 53 | 54 | // Convert from compressed string representation of encoded mask. 55 | void rleFrString( RLE *R, char *s, siz h, siz w ); 56 | -------------------------------------------------------------------------------- /lib/setup.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) Facebook, Inc. and its affiliates. All Rights Reserved. 2 | #!/usr/bin/env python 3 | 4 | import glob 5 | import os 6 | 7 | import torch 8 | from setuptools import find_packages 9 | from setuptools import setup 10 | from torch.utils.cpp_extension import CUDA_HOME 11 | from torch.utils.cpp_extension import CppExtension 12 | from torch.utils.cpp_extension import CUDAExtension 13 | 14 | requirements = ["torch", "torchvision"] 15 | 16 | 17 | def get_extensions(): 18 | this_dir = os.path.dirname(os.path.abspath(__file__)) 19 | extensions_dir = os.path.join(this_dir, "model", "csrc") 20 | 21 | main_file = glob.glob(os.path.join(extensions_dir, "*.cpp")) 22 | source_cpu = glob.glob(os.path.join(extensions_dir, "cpu", "*.cpp")) 23 | source_cuda = glob.glob(os.path.join(extensions_dir, "cuda", "*.cu")) 24 | 25 | sources = main_file + source_cpu 26 | extension = CppExtension 27 | 28 | extra_compile_args = {"cxx": []} 29 | define_macros = [] 30 | 31 | if torch.cuda.is_available() and CUDA_HOME is not None: 32 | extension = CUDAExtension 33 | sources += source_cuda 34 | define_macros += [("WITH_CUDA", None)] 35 | extra_compile_args["nvcc"] = [ 36 | "-DCUDA_HAS_FP16=1", 37 | "-D__CUDA_NO_HALF_OPERATORS__", 38 | "-D__CUDA_NO_HALF_CONVERSIONS__", 39 | "-D__CUDA_NO_HALF2_OPERATORS__", 40 | ] 41 | 42 | sources = [os.path.join(extensions_dir, s) for s in sources] 43 | 44 | include_dirs = [extensions_dir] 45 | 46 | ext_modules = [ 47 | extension( 48 | "model._C", 49 | sources, 50 | include_dirs=include_dirs, 51 | define_macros=define_macros, 52 | extra_compile_args=extra_compile_args, 53 | ) 54 | ] 55 | 56 | return ext_modules 57 | 58 | 59 | setup( 60 | name="faster_rcnn", 61 | version="0.1", 62 | description="object detection in pytorch", 63 | packages=find_packages(exclude=("configs", "tests",)), 64 | # install_requires=requirements, 65 | ext_modules=get_extensions(), 66 | cmdclass={"build_ext": torch.utils.cpp_extension.BuildExtension}, 67 | ) 68 | -------------------------------------------------------------------------------- /lib/model/roi_align/functions/roi_align.py: -------------------------------------------------------------------------------- 1 | import torch 2 | from torch.autograd import Function 3 | from .._ext import roi_align 4 | 5 | 6 | # TODO use save_for_backward instead 7 | class RoIAlignFunction(Function): 8 | def __init__(self, aligned_height, aligned_width, spatial_scale): 9 | self.aligned_width = int(aligned_width) 10 | self.aligned_height = int(aligned_height) 11 | self.spatial_scale = float(spatial_scale) 12 | self.rois = None 13 | self.feature_size = None 14 | 15 | def forward(self, features, rois): 16 | self.rois = rois 17 | self.feature_size = features.size() 18 | 19 | batch_size, num_channels, data_height, data_width = features.size() 20 | num_rois = rois.size(0) 21 | 22 | output = features.new(num_rois, num_channels, self.aligned_height, self.aligned_width).zero_() 23 | if features.is_cuda: 24 | roi_align.roi_align_forward_cuda(self.aligned_height, 25 | self.aligned_width, 26 | self.spatial_scale, features, 27 | rois, output) 28 | else: 29 | roi_align.roi_align_forward(self.aligned_height, 30 | self.aligned_width, 31 | self.spatial_scale, features, 32 | rois, output) 33 | # raise NotImplementedError 34 | 35 | return output 36 | 37 | def backward(self, grad_output): 38 | assert(self.feature_size is not None and grad_output.is_cuda) 39 | 40 | batch_size, num_channels, data_height, data_width = self.feature_size 41 | 42 | grad_input = self.rois.new(batch_size, num_channels, data_height, 43 | data_width).zero_() 44 | roi_align.roi_align_backward_cuda(self.aligned_height, 45 | self.aligned_width, 46 | self.spatial_scale, grad_output, 47 | self.rois, grad_input) 48 | 49 | # print grad_input 50 | 51 | return grad_input, None 52 | -------------------------------------------------------------------------------- /lib/model/faster_rcnn/vgg16.py: -------------------------------------------------------------------------------- 1 | # -------------------------------------------------------- 2 | # Tensorflow Faster R-CNN 3 | # Licensed under The MIT License [see LICENSE for details] 4 | # Written by Xinlei Chen 5 | # -------------------------------------------------------- 6 | from __future__ import absolute_import 7 | from __future__ import division 8 | from __future__ import print_function 9 | 10 | import torch 11 | import torch.nn as nn 12 | import torch.nn.functional as F 13 | from torch.autograd import Variable 14 | import math 15 | import torchvision.models as models 16 | from model.faster_rcnn.faster_rcnn import _fasterRCNN 17 | import pdb 18 | 19 | class vgg16(_fasterRCNN): 20 | def __init__(self, classes, pretrained=False, class_agnostic=False): 21 | self.model_path = 'data/pretrained_model/vgg16_caffe.pth' 22 | self.dout_base_model = 512 23 | self.pretrained = pretrained 24 | self.class_agnostic = class_agnostic 25 | 26 | _fasterRCNN.__init__(self, classes, class_agnostic) 27 | 28 | def _init_modules(self): 29 | vgg = models.vgg16() 30 | if self.pretrained: 31 | print("Loading pretrained weights from %s" %(self.model_path)) 32 | state_dict = torch.load(self.model_path) 33 | vgg.load_state_dict({k:v for k,v in state_dict.items() if k in vgg.state_dict()}) 34 | 35 | vgg.classifier = nn.Sequential(*list(vgg.classifier._modules.values())[:-1]) 36 | 37 | # not using the last maxpool layer 38 | self.RCNN_base = nn.Sequential(*list(vgg.features._modules.values())[:-1]) 39 | 40 | # Fix the layers before conv3: 41 | for layer in range(10): 42 | for p in self.RCNN_base[layer].parameters(): p.requires_grad = False 43 | 44 | # self.RCNN_base = _RCNN_base(vgg.features, self.classes, self.dout_base_model) 45 | 46 | self.RCNN_top = vgg.classifier 47 | 48 | # not using the last maxpool layer 49 | self.RCNN_cls_score = nn.Linear(4096, self.n_classes) 50 | 51 | if self.class_agnostic: 52 | self.RCNN_bbox_pred = nn.Linear(4096, 4) 53 | else: 54 | self.RCNN_bbox_pred = nn.Linear(4096, 4 * self.n_classes) 55 | 56 | def _head_to_tail(self, pool5): 57 | 58 | pool5_flat = pool5.view(pool5.size(0), -1) 59 | fc7 = self.RCNN_top(pool5_flat) 60 | 61 | return fc7 62 | 63 | -------------------------------------------------------------------------------- /lib/model/csrc/cuda/vision.h: -------------------------------------------------------------------------------- 1 | // Copyright (c) Facebook, Inc. and its affiliates. All Rights Reserved. 2 | #pragma once 3 | #include 4 | 5 | 6 | at::Tensor ROIAlign_forward_cuda(const at::Tensor& input, 7 | const at::Tensor& rois, 8 | const float spatial_scale, 9 | const int pooled_height, 10 | const int pooled_width, 11 | const int sampling_ratio); 12 | 13 | at::Tensor ROIAlign_backward_cuda(const at::Tensor& grad, 14 | const at::Tensor& rois, 15 | const float spatial_scale, 16 | const int pooled_height, 17 | const int pooled_width, 18 | const int batch_size, 19 | const int channels, 20 | const int height, 21 | const int width, 22 | const int sampling_ratio); 23 | 24 | 25 | std::tuple ROIPool_forward_cuda(const at::Tensor& input, 26 | const at::Tensor& rois, 27 | const float spatial_scale, 28 | const int pooled_height, 29 | const int pooled_width); 30 | 31 | at::Tensor ROIPool_backward_cuda(const at::Tensor& grad, 32 | const at::Tensor& input, 33 | const at::Tensor& rois, 34 | const at::Tensor& argmax, 35 | const float spatial_scale, 36 | const int pooled_height, 37 | const int pooled_width, 38 | const int batch_size, 39 | const int channels, 40 | const int height, 41 | const int width); 42 | 43 | at::Tensor nms_cuda(const at::Tensor boxes, float nms_overlap_thresh); 44 | 45 | 46 | at::Tensor compute_flow_cuda(const at::Tensor& boxes, 47 | const int height, 48 | const int width); 49 | -------------------------------------------------------------------------------- /lib/model/roi_layers/roi_align.py: -------------------------------------------------------------------------------- 1 | # Copyright (c) Facebook, Inc. and its affiliates. All Rights Reserved. 2 | import torch 3 | from torch import nn 4 | from torch.autograd import Function 5 | from torch.autograd.function import once_differentiable 6 | from torch.nn.modules.utils import _pair 7 | 8 | from model import _C 9 | 10 | import pdb 11 | 12 | class _ROIAlign(Function): 13 | @staticmethod 14 | def forward(ctx, input, roi, output_size, spatial_scale, sampling_ratio): 15 | ctx.save_for_backward(roi) 16 | ctx.output_size = _pair(output_size) 17 | ctx.spatial_scale = spatial_scale 18 | ctx.sampling_ratio = sampling_ratio 19 | ctx.input_shape = input.size() 20 | output = _C.roi_align_forward(input, roi, spatial_scale, output_size[0], output_size[1], sampling_ratio) 21 | return output 22 | 23 | @staticmethod 24 | @once_differentiable 25 | def backward(ctx, grad_output): 26 | rois, = ctx.saved_tensors 27 | output_size = ctx.output_size 28 | spatial_scale = ctx.spatial_scale 29 | sampling_ratio = ctx.sampling_ratio 30 | bs, ch, h, w = ctx.input_shape 31 | grad_input = _C.roi_align_backward( 32 | grad_output, 33 | rois, 34 | spatial_scale, 35 | output_size[0], 36 | output_size[1], 37 | bs, 38 | ch, 39 | h, 40 | w, 41 | sampling_ratio, 42 | ) 43 | return grad_input, None, None, None, None 44 | 45 | 46 | roi_align = _ROIAlign.apply 47 | 48 | 49 | class ROIAlign(nn.Module): 50 | def __init__(self, output_size, spatial_scale, sampling_ratio): 51 | super(ROIAlign, self).__init__() 52 | self.output_size = output_size 53 | self.spatial_scale = spatial_scale 54 | self.sampling_ratio = sampling_ratio 55 | 56 | def forward(self, input, rois): 57 | return roi_align( 58 | input, rois, self.output_size, self.spatial_scale, self.sampling_ratio 59 | ) 60 | 61 | def __repr__(self): 62 | tmpstr = self.__class__.__name__ + "(" 63 | tmpstr += "output_size=" + str(self.output_size) 64 | tmpstr += ", spatial_scale=" + str(self.spatial_scale) 65 | tmpstr += ", sampling_ratio=" + str(self.sampling_ratio) 66 | tmpstr += ")" 67 | return tmpstr 68 | -------------------------------------------------------------------------------- /lib/model/roi_crop/functions/gridgen.py: -------------------------------------------------------------------------------- 1 | # functions/add.py 2 | import torch 3 | from torch.autograd import Function 4 | import numpy as np 5 | 6 | 7 | class AffineGridGenFunction(Function): 8 | def __init__(self, height, width,lr=1): 9 | super(AffineGridGenFunction, self).__init__() 10 | self.lr = lr 11 | self.height, self.width = height, width 12 | self.grid = np.zeros( [self.height, self.width, 3], dtype=np.float32) 13 | self.grid[:,:,0] = np.expand_dims(np.repeat(np.expand_dims(np.arange(-1, 1, 2.0/(self.height)), 0), repeats = self.width, axis = 0).T, 0) 14 | self.grid[:,:,1] = np.expand_dims(np.repeat(np.expand_dims(np.arange(-1, 1, 2.0/(self.width)), 0), repeats = self.height, axis = 0), 0) 15 | # self.grid[:,:,0] = np.expand_dims(np.repeat(np.expand_dims(np.arange(-1, 1, 2.0/(self.height - 1)), 0), repeats = self.width, axis = 0).T, 0) 16 | # self.grid[:,:,1] = np.expand_dims(np.repeat(np.expand_dims(np.arange(-1, 1, 2.0/(self.width - 1)), 0), repeats = self.height, axis = 0), 0) 17 | self.grid[:,:,2] = np.ones([self.height, width]) 18 | self.grid = torch.from_numpy(self.grid.astype(np.float32)) 19 | #print(self.grid) 20 | 21 | def forward(self, input1): 22 | self.input1 = input1 23 | output = input1.new(torch.Size([input1.size(0)]) + self.grid.size()).zero_() 24 | self.batchgrid = input1.new(torch.Size([input1.size(0)]) + self.grid.size()).zero_() 25 | for i in range(input1.size(0)): 26 | self.batchgrid[i] = self.grid.astype(self.batchgrid[i]) 27 | 28 | # if input1.is_cuda: 29 | # self.batchgrid = self.batchgrid.cuda() 30 | # output = output.cuda() 31 | 32 | for i in range(input1.size(0)): 33 | output = torch.bmm(self.batchgrid.view(-1, self.height*self.width, 3), torch.transpose(input1, 1, 2)).view(-1, self.height, self.width, 2) 34 | 35 | return output 36 | 37 | def backward(self, grad_output): 38 | 39 | grad_input1 = self.input1.new(self.input1.size()).zero_() 40 | 41 | # if grad_output.is_cuda: 42 | # self.batchgrid = self.batchgrid.cuda() 43 | # grad_input1 = grad_input1.cuda() 44 | 45 | grad_input1 = torch.baddbmm(grad_input1, torch.transpose(grad_output.view(-1, self.height*self.width, 2), 1,2), self.batchgrid.view(-1, self.height*self.width, 3)) 46 | return grad_input1 47 | -------------------------------------------------------------------------------- /lib/model/roi_align/src/roi_align_cuda.c: -------------------------------------------------------------------------------- 1 | #include 2 | #include 3 | #include "roi_align_kernel.h" 4 | 5 | extern THCState *state; 6 | 7 | int roi_align_forward_cuda(int aligned_height, int aligned_width, float spatial_scale, 8 | THCudaTensor * features, THCudaTensor * rois, THCudaTensor * output) 9 | { 10 | // Grab the input tensor 11 | float * data_flat = THCudaTensor_data(state, features); 12 | float * rois_flat = THCudaTensor_data(state, rois); 13 | 14 | float * output_flat = THCudaTensor_data(state, output); 15 | 16 | // Number of ROIs 17 | int num_rois = THCudaTensor_size(state, rois, 0); 18 | int size_rois = THCudaTensor_size(state, rois, 1); 19 | if (size_rois != 5) 20 | { 21 | return 0; 22 | } 23 | 24 | // data height 25 | int data_height = THCudaTensor_size(state, features, 2); 26 | // data width 27 | int data_width = THCudaTensor_size(state, features, 3); 28 | // Number of channels 29 | int num_channels = THCudaTensor_size(state, features, 1); 30 | 31 | cudaStream_t stream = THCState_getCurrentStream(state); 32 | 33 | ROIAlignForwardLaucher( 34 | data_flat, spatial_scale, num_rois, data_height, 35 | data_width, num_channels, aligned_height, 36 | aligned_width, rois_flat, 37 | output_flat, stream); 38 | 39 | return 1; 40 | } 41 | 42 | int roi_align_backward_cuda(int aligned_height, int aligned_width, float spatial_scale, 43 | THCudaTensor * top_grad, THCudaTensor * rois, THCudaTensor * bottom_grad) 44 | { 45 | // Grab the input tensor 46 | float * top_grad_flat = THCudaTensor_data(state, top_grad); 47 | float * rois_flat = THCudaTensor_data(state, rois); 48 | 49 | float * bottom_grad_flat = THCudaTensor_data(state, bottom_grad); 50 | 51 | // Number of ROIs 52 | int num_rois = THCudaTensor_size(state, rois, 0); 53 | int size_rois = THCudaTensor_size(state, rois, 1); 54 | if (size_rois != 5) 55 | { 56 | return 0; 57 | } 58 | 59 | // batch size 60 | int batch_size = THCudaTensor_size(state, bottom_grad, 0); 61 | // data height 62 | int data_height = THCudaTensor_size(state, bottom_grad, 2); 63 | // data width 64 | int data_width = THCudaTensor_size(state, bottom_grad, 3); 65 | // Number of channels 66 | int num_channels = THCudaTensor_size(state, bottom_grad, 1); 67 | 68 | cudaStream_t stream = THCState_getCurrentStream(state); 69 | ROIAlignBackwardLaucher( 70 | top_grad_flat, spatial_scale, batch_size, num_rois, data_height, 71 | data_width, num_channels, aligned_height, 72 | aligned_width, rois_flat, 73 | bottom_grad_flat, stream); 74 | 75 | return 1; 76 | } 77 | -------------------------------------------------------------------------------- /lib/model/utils/logger.py: -------------------------------------------------------------------------------- 1 | # Code referenced from https://gist.github.com/gyglim/1f8dfb1b5c82627ae3efcfbbadb9f514 2 | import tensorflow as tf 3 | import numpy as np 4 | import scipy.misc 5 | try: 6 | from StringIO import StringIO # Python 2.7 7 | except ImportError: 8 | from io import BytesIO # Python 3.x 9 | 10 | 11 | class Logger(object): 12 | 13 | def __init__(self, log_dir): 14 | """Create a summary writer logging to log_dir.""" 15 | self.writer = tf.summary.FileWriter(log_dir) 16 | 17 | def scalar_summary(self, tag, value, step): 18 | """Log a scalar variable.""" 19 | summary = tf.Summary(value=[tf.Summary.Value(tag=tag, simple_value=value)]) 20 | self.writer.add_summary(summary, step) 21 | 22 | def image_summary(self, tag, images, step): 23 | """Log a list of images.""" 24 | 25 | img_summaries = [] 26 | for i, img in enumerate(images): 27 | # Write the image to a string 28 | try: 29 | s = StringIO() 30 | except: 31 | s = BytesIO() 32 | scipy.misc.toimage(img).save(s, format="png") 33 | 34 | # Create an Image object 35 | img_sum = tf.Summary.Image(encoded_image_string=s.getvalue(), 36 | height=img.shape[0], 37 | width=img.shape[1]) 38 | # Create a Summary value 39 | img_summaries.append(tf.Summary.Value(tag='%s/%d' % (tag, i), image=img_sum)) 40 | 41 | # Create and write Summary 42 | summary = tf.Summary(value=img_summaries) 43 | self.writer.add_summary(summary, step) 44 | 45 | def histo_summary(self, tag, values, step, bins=): 46 | """Log a histogram of the tensor of values.""" 47 | 48 | # Create a histogram using numpy 49 | counts, bin_edges = np.histogram(values, bins=bins) 50 | 51 | # Fill the fields of the histogram proto 52 | hist = tf.HistogramProto() 53 | hist.min = float(np.min(values)) 54 | hist.max = float(np.max(values)) 55 | hist.num = int(np.prod(values.shape)) 56 | hist.sum = float(np.sum(values)) 57 | hist.sum_squares = float(np.sum(values**2)) 58 | 59 | # Drop the start of the first bin 60 | bin_edges = bin_edges[1:] 61 | 62 | # Add bin edges and counts 63 | for edge in bin_edges: 64 | hist.bucket_limit.append(edge) 65 | for c in counts: 66 | hist.bucket.append(c) 67 | 68 | # Create and write Summary 69 | summary = tf.Summary(value=[tf.Summary.Value(tag=tag, histo=hist)]) 70 | self.writer.add_summary(summary, step) 71 | self.writer.flush() 72 | -------------------------------------------------------------------------------- /lib/model/csrc/cpu/nms_cpu.cpp: -------------------------------------------------------------------------------- 1 | // Copyright (c) Facebook, Inc. and its affiliates. All Rights Reserved. 2 | #include "cpu/vision.h" 3 | 4 | 5 | template 6 | at::Tensor nms_cpu_kernel(const at::Tensor& dets, 7 | const at::Tensor& scores, 8 | const float threshold) { 9 | AT_ASSERTM(!dets.type().is_cuda(), "dets must be a CPU tensor"); 10 | AT_ASSERTM(!scores.type().is_cuda(), "scores must be a CPU tensor"); 11 | AT_ASSERTM(dets.type() == scores.type(), "dets should have the same type as scores"); 12 | 13 | if (dets.numel() == 0) { 14 | return at::empty({0}, dets.options().dtype(at::kLong).device(at::kCPU)); 15 | } 16 | 17 | auto x1_t = dets.select(1, 0).contiguous(); 18 | auto y1_t = dets.select(1, 1).contiguous(); 19 | auto x2_t = dets.select(1, 2).contiguous(); 20 | auto y2_t = dets.select(1, 3).contiguous(); 21 | 22 | at::Tensor areas_t = (x2_t - x1_t + 1) * (y2_t - y1_t + 1); 23 | 24 | auto order_t = std::get<1>(scores.sort(0, /* descending=*/true)); 25 | 26 | auto ndets = dets.size(0); 27 | at::Tensor suppressed_t = at::zeros({ndets}, dets.options().dtype(at::kByte).device(at::kCPU)); 28 | 29 | auto suppressed = suppressed_t.data(); 30 | auto order = order_t.data(); 31 | auto x1 = x1_t.data(); 32 | auto y1 = y1_t.data(); 33 | auto x2 = x2_t.data(); 34 | auto y2 = y2_t.data(); 35 | auto areas = areas_t.data(); 36 | 37 | for (int64_t _i = 0; _i < ndets; _i++) { 38 | auto i = order[_i]; 39 | if (suppressed[i] == 1) 40 | continue; 41 | auto ix1 = x1[i]; 42 | auto iy1 = y1[i]; 43 | auto ix2 = x2[i]; 44 | auto iy2 = y2[i]; 45 | auto iarea = areas[i]; 46 | 47 | for (int64_t _j = _i + 1; _j < ndets; _j++) { 48 | auto j = order[_j]; 49 | if (suppressed[j] == 1) 50 | continue; 51 | auto xx1 = std::max(ix1, x1[j]); 52 | auto yy1 = std::max(iy1, y1[j]); 53 | auto xx2 = std::min(ix2, x2[j]); 54 | auto yy2 = std::min(iy2, y2[j]); 55 | 56 | auto w = std::max(static_cast(0), xx2 - xx1 + 1); 57 | auto h = std::max(static_cast(0), yy2 - yy1 + 1); 58 | auto inter = w * h; 59 | auto ovr = inter / (iarea + areas[j] - inter); 60 | if (ovr >= threshold) 61 | suppressed[j] = 1; 62 | } 63 | } 64 | return at::nonzero(suppressed_t == 0).squeeze(1); 65 | } 66 | 67 | at::Tensor nms_cpu(const at::Tensor& dets, 68 | const at::Tensor& scores, 69 | const float threshold) { 70 | at::Tensor result; 71 | AT_DISPATCH_FLOATING_TYPES(dets.type(), "nms", [&] { 72 | result = nms_cpu_kernel(dets, scores, threshold); 73 | }); 74 | return result; 75 | } 76 | -------------------------------------------------------------------------------- /lib/model/roi_crop/src/roi_crop_cuda_kernel.h: -------------------------------------------------------------------------------- 1 | #ifdef __cplusplus 2 | extern "C" { 3 | #endif 4 | 5 | 6 | int BilinearSamplerBHWD_updateOutput_cuda_kernel(/*output->size[3]*/int oc, 7 | /*output->size[2]*/int ow, 8 | /*output->size[1]*/int oh, 9 | /*output->size[0]*/int ob, 10 | /*THCudaTensor_size(state, inputImages, 3)*/int ic, 11 | /*THCudaTensor_size(state, inputImages, 1)*/int ih, 12 | /*THCudaTensor_size(state, inputImages, 2)*/int iw, 13 | /*THCudaTensor_size(state, inputImages, 0)*/int ib, 14 | /*THCudaTensor *inputImages*/float *inputImages, int isb, int isc, int ish, int isw, 15 | /*THCudaTensor *grids*/float *grids, int gsb, int gsc, int gsh, int gsw, 16 | /*THCudaTensor *output*/float *output, int osb, int osc, int osh, int osw, 17 | /*THCState_getCurrentStream(state)*/cudaStream_t stream); 18 | 19 | int BilinearSamplerBHWD_updateGradInput_cuda_kernel(/*gradOutput->size[3]*/int goc, 20 | /*gradOutput->size[2]*/int gow, 21 | /*gradOutput->size[1]*/int goh, 22 | /*gradOutput->size[0]*/int gob, 23 | /*THCudaTensor_size(state, inputImages, 3)*/int ic, 24 | /*THCudaTensor_size(state, inputImages, 1)*/int ih, 25 | /*THCudaTensor_size(state, inputImages, 2)*/int iw, 26 | /*THCudaTensor_size(state, inputImages, 0)*/int ib, 27 | /*THCudaTensor *inputImages*/float *inputImages, int isb, int isc, int ish, int isw, 28 | /*THCudaTensor *grids*/float *grids, int gsb, int gsc, int gsh, int gsw, 29 | /*THCudaTensor *gradInputImages*/float *gradInputImages, int gisb, int gisc, int gish, int gisw, 30 | /*THCudaTensor *gradGrids*/float *gradGrids, int ggsb, int ggsc, int ggsh, int ggsw, 31 | /*THCudaTensor *gradOutput*/float *gradOutput, int gosb, int gosc, int gosh, int gosw, 32 | /*THCState_getCurrentStream(state)*/cudaStream_t stream); 33 | 34 | 35 | #ifdef __cplusplus 36 | } 37 | #endif 38 | -------------------------------------------------------------------------------- /lib/datasets/factory.py: -------------------------------------------------------------------------------- 1 | # -------------------------------------------------------- 2 | # Fast R-CNN 3 | # Copyright (c) 2015 Microsoft 4 | # Licensed under The MIT License [see LICENSE for details] 5 | # Written by Ross Girshick 6 | # -------------------------------------------------------- 7 | 8 | """Factory method for easily getting imdbs by name.""" 9 | from __future__ import absolute_import 10 | from __future__ import division 11 | from __future__ import print_function 12 | 13 | __sets = {} 14 | from datasets.pascal_voc import pascal_voc 15 | # from datasets.coco import coco 16 | from datasets.imagenet import imagenet 17 | from datasets.vg import vg 18 | 19 | import numpy as np 20 | 21 | # Set up voc__ 22 | for year in ['2007', '2012']: 23 | for split in ['train', 'val', 'trainval', 'test']: 24 | name = 'voc_{}_{}'.format(year, split) 25 | __sets[name] = (lambda split=split, year=year: pascal_voc(split, year)) 26 | 27 | # Set up coco_2014_ 28 | for year in ['2014']: 29 | for split in ['train', 'val', 'minival', 'valminusminival', 'trainval']: 30 | name = 'coco_{}_{}'.format(year, split) 31 | __sets[name] = (lambda split=split, year=year: coco(split, year)) 32 | 33 | # Set up coco_2014_cap_ 34 | for year in ['2014']: 35 | for split in ['train', 'val', 'capval', 'valminuscapval', 'trainval']: 36 | name = 'coco_{}_{}'.format(year, split) 37 | __sets[name] = (lambda split=split, year=year: coco(split, year)) 38 | 39 | # Set up coco_2015_ 40 | for year in ['2015']: 41 | for split in ['test', 'test-dev']: 42 | name = 'coco_{}_{}'.format(year, split) 43 | __sets[name] = (lambda split=split, year=year: coco(split, year)) 44 | 45 | # Set up vg_ 46 | # for version in ['1600-400-20']: 47 | # for split in ['minitrain', 'train', 'minival', 'val', 'test']: 48 | # name = 'vg_{}_{}'.format(version,split) 49 | # __sets[name] = (lambda split=split, version=version: vg(version, split)) 50 | for version in ['150-50-20', '150-50-50', '500-150-80', '750-250-150', '1750-700-450', '1600-400-20']: 51 | for split in ['minitrain', 'smalltrain', 'train', 'minival', 'smallval', 'val', 'test']: 52 | name = 'vg_{}_{}'.format(version,split) 53 | __sets[name] = (lambda split=split, version=version: vg(version, split)) 54 | 55 | # set up image net. 56 | for split in ['train', 'val', 'val1', 'val2', 'test']: 57 | name = 'imagenet_{}'.format(split) 58 | devkit_path = 'data/imagenet/ILSVRC/devkit' 59 | data_path = 'data/imagenet/ILSVRC' 60 | __sets[name] = (lambda split=split, devkit_path=devkit_path, data_path=data_path: imagenet(split,devkit_path,data_path)) 61 | 62 | def get_imdb(name): 63 | """Get an imdb (image database) by name.""" 64 | if name not in __sets: 65 | raise KeyError('Unknown dataset: {}'.format(name)) 66 | return __sets[name]() 67 | 68 | 69 | def list_imdbs(): 70 | """List all registered imdbs.""" 71 | return list(__sets.keys()) 72 | -------------------------------------------------------------------------------- /lib/model/roi_pooling/src/roi_pooling_cuda.c: -------------------------------------------------------------------------------- 1 | #include 2 | #include 3 | #include "roi_pooling_kernel.h" 4 | 5 | extern THCState *state; 6 | 7 | int roi_pooling_forward_cuda(int pooled_height, int pooled_width, float spatial_scale, 8 | THCudaTensor * features, THCudaTensor * rois, THCudaTensor * output, THCudaIntTensor * argmax) 9 | { 10 | // Grab the input tensor 11 | float * data_flat = THCudaTensor_data(state, features); 12 | float * rois_flat = THCudaTensor_data(state, rois); 13 | 14 | float * output_flat = THCudaTensor_data(state, output); 15 | int * argmax_flat = THCudaIntTensor_data(state, argmax); 16 | 17 | // Number of ROIs 18 | int num_rois = THCudaTensor_size(state, rois, 0); 19 | int size_rois = THCudaTensor_size(state, rois, 1); 20 | if (size_rois != 5) 21 | { 22 | return 0; 23 | } 24 | 25 | // batch size 26 | // int batch_size = THCudaTensor_size(state, features, 0); 27 | // if (batch_size != 1) 28 | // { 29 | // return 0; 30 | // } 31 | // data height 32 | int data_height = THCudaTensor_size(state, features, 2); 33 | // data width 34 | int data_width = THCudaTensor_size(state, features, 3); 35 | // Number of channels 36 | int num_channels = THCudaTensor_size(state, features, 1); 37 | 38 | cudaStream_t stream = THCState_getCurrentStream(state); 39 | 40 | ROIPoolForwardLaucher( 41 | data_flat, spatial_scale, num_rois, data_height, 42 | data_width, num_channels, pooled_height, 43 | pooled_width, rois_flat, 44 | output_flat, argmax_flat, stream); 45 | 46 | return 1; 47 | } 48 | 49 | int roi_pooling_backward_cuda(int pooled_height, int pooled_width, float spatial_scale, 50 | THCudaTensor * top_grad, THCudaTensor * rois, THCudaTensor * bottom_grad, THCudaIntTensor * argmax) 51 | { 52 | // Grab the input tensor 53 | float * top_grad_flat = THCudaTensor_data(state, top_grad); 54 | float * rois_flat = THCudaTensor_data(state, rois); 55 | 56 | float * bottom_grad_flat = THCudaTensor_data(state, bottom_grad); 57 | int * argmax_flat = THCudaIntTensor_data(state, argmax); 58 | 59 | // Number of ROIs 60 | int num_rois = THCudaTensor_size(state, rois, 0); 61 | int size_rois = THCudaTensor_size(state, rois, 1); 62 | if (size_rois != 5) 63 | { 64 | return 0; 65 | } 66 | 67 | // batch size 68 | int batch_size = THCudaTensor_size(state, bottom_grad, 0); 69 | // if (batch_size != 1) 70 | // { 71 | // return 0; 72 | // } 73 | // data height 74 | int data_height = THCudaTensor_size(state, bottom_grad, 2); 75 | // data width 76 | int data_width = THCudaTensor_size(state, bottom_grad, 3); 77 | // Number of channels 78 | int num_channels = THCudaTensor_size(state, bottom_grad, 1); 79 | 80 | cudaStream_t stream = THCState_getCurrentStream(state); 81 | ROIPoolBackwardLaucher( 82 | top_grad_flat, spatial_scale, batch_size, num_rois, data_height, 83 | data_width, num_channels, pooled_height, 84 | pooled_width, rois_flat, 85 | bottom_grad_flat, argmax_flat, stream); 86 | 87 | return 1; 88 | } 89 | -------------------------------------------------------------------------------- /lib/model/utils/viz_hand_obj.py: -------------------------------------------------------------------------------- 1 | import os, json, glob, random 2 | import numpy as np 3 | from tqdm import tqdm 4 | from PIL import Image, ImageDraw, ImageFont 5 | random.seed(0) 6 | 7 | def ratio2coord(ratio, width, height): 8 | """ 9 | ratio = [x1, y1, x2, y2] 10 | return image infos 11 | """ 12 | 13 | x1, y1, x2, y2 = int(float(ratio[0])*width), int(float(ratio[1])*height), int(float(ratio[2])*width), int(float(ratio[3])*height) 14 | 15 | x1 = max(x1, 0) 16 | y1 = max(y1, 0) 17 | x2 = min(x2, width) 18 | y2 = min(y2, height) 19 | 20 | bbox = [x1, y1, x2, y2] 21 | 22 | return bbox 23 | 24 | def bbox2center(bbox): 25 | return (int((bbox[0]+bbox[2])/2), int((bbox[1]+bbox[3])/2)) 26 | 27 | 28 | def draw_obj_mask(image, draw, obj_idx, obj_bbox, obj_score, width, height, font): 29 | 30 | mask = Image.new('RGBA', (width, height)) 31 | pmask = ImageDraw.Draw(mask) 32 | pmask.rectangle(obj_bbox, outline=obj_rgb, width=4, fill=obj_rgba) 33 | image.paste(mask, (0,0), mask) 34 | 35 | draw.rectangle([obj_bbox[0], max(0, obj_bbox[1]-30), obj_bbox[0]+32, max(0, obj_bbox[1]-30)+30], fill=(255, 255, 255), outline=obj_rgb, width=4) 36 | draw.text((obj_bbox[0]+5, max(0, obj_bbox[1]-30)-2), f'O', font=font, fill=(0,0,0)) # 37 | 38 | return image 39 | 40 | 41 | def draw_hand_mask(image, draw, hand_idx, hand_bbox, hand_score, side, state, width, height, font): 42 | 43 | if side == 0: 44 | side_idx = 0 45 | elif side == 1: 46 | side_idx = 1 47 | mask = Image.new('RGBA', (width, height)) 48 | pmask = ImageDraw.Draw(mask) 49 | pmask.rectangle(hand_bbox, outline=hand_rgb[side_idx], width=4, fill=hand_rgba[side_idx]) 50 | image.paste(mask, (0,0), mask) 51 | # text 52 | 53 | draw = ImageDraw.Draw(image) 54 | draw.rectangle([hand_bbox[0], max(0, hand_bbox[1]-30), hand_bbox[0]+62, max(0, hand_bbox[1]-30)+30], fill=(255, 255, 255), outline=hand_rgb[side_idx], width=4) 55 | draw.text((hand_bbox[0]+6, max(0, hand_bbox[1]-30)-2), f'{side_map3[int(float(side))]}-{state_map2[int(float(state))]}', font=font, fill=(0,0,0)) # 56 | 57 | return image 58 | 59 | def draw_line_point(draw, side_idx, hand_center, object_center): 60 | 61 | draw.line([hand_center, object_center], fill=hand_rgb[side_idx], width=4) 62 | x, y = hand_center[0], hand_center[1] 63 | r=7 64 | draw.ellipse((x-r, y-r, x+r, y+r), fill=hand_rgb[side_idx]) 65 | x, y = object_center[0], object_center[1] 66 | draw.ellipse((x-r, y-r, x+r, y+r), fill=obj_rgb) 67 | 68 | 69 | color_rgb = [(255,255,0), (255, 128,0), (128,255,0), (0,128,255), (0,0,255), (127,0,255), (255,0,255), (255,0,127), (255,0,0), (255,204,153), (255,102,102), (153,255,153), (153,153,255), (0,0,153)] 70 | color_rgba = [(255,255,0,70), (255, 128,0,70), (128,255,0,70), (0,128,255,70), (0,0,255,70), (127,0,255,70), (255,0,255,70), (255,0,127,70), (255,0,0,70), (255,204,153,70), (255,102,102,70), (153,255,153,70), (153,153,255,70), (0,0,153,70)] 71 | 72 | 73 | hand_rgb = [(0, 90, 181), (220, 50, 32)] 74 | hand_rgba = [(0, 90, 181, 70), (220, 50, 32, 70)] 75 | 76 | obj_rgb = (255, 194, 10) 77 | obj_rgba = (255, 194, 10, 70) 78 | 79 | 80 | side_map = {'l':'Left', 'r':'Right'} 81 | side_map2 = {0:'Left', 1:'Right'} 82 | side_map3 = {0:'L', 1:'R'} 83 | state_map = {0:'No Contact', 1:'Self Contact', 2:'Another Person', 3:'Portable Object', 4:'Stationary Object'} 84 | state_map2 = {0:'N', 1:'S', 2:'O', 3:'P', 4:'F'} 85 | 86 | -------------------------------------------------------------------------------- /lib/roi_data_layer/minibatch.py: -------------------------------------------------------------------------------- 1 | # -------------------------------------------------------- 2 | # Fast R-CNN 3 | # Copyright (c) 2015 Microsoft 4 | # Licensed under The MIT License [see LICENSE for details] 5 | # Written by Ross Girshick and Xinlei Chen 6 | # -------------------------------------------------------- 7 | 8 | """Compute minibatch blobs for training a Fast R-CNN network.""" 9 | from __future__ import absolute_import 10 | from __future__ import division 11 | from __future__ import print_function 12 | 13 | import cv2 14 | import numpy as np 15 | import numpy.random as npr 16 | # from scipy.misc import imread 17 | from model.utils.config import cfg 18 | from model.utils.blob import prep_im_for_blob, im_list_to_blob 19 | import pdb 20 | def get_minibatch(roidb, num_classes): 21 | """Given a roidb, construct a minibatch sampled from it.""" 22 | num_images = len(roidb) 23 | # Sample random scales to use for each image in this batch 24 | random_scale_inds = npr.randint(0, high=len(cfg.TRAIN.SCALES), 25 | size=num_images) 26 | assert(cfg.TRAIN.BATCH_SIZE % num_images == 0), \ 27 | 'num_images ({}) must divide BATCH_SIZE ({})'. \ 28 | format(num_images, cfg.TRAIN.BATCH_SIZE) 29 | 30 | # Get the input image blob, formatted for caffe 31 | im_blob, im_scales = _get_image_blob(roidb, random_scale_inds) 32 | 33 | blobs = {'data': im_blob} 34 | 35 | assert len(im_scales) == 1, "Single batch only" 36 | assert len(roidb) == 1, "Single batch only" 37 | 38 | # gt boxes: (x1, y1, x2, y2, cls) 39 | if cfg.TRAIN.USE_ALL_GT: 40 | # Include all ground truth boxes 41 | gt_inds = np.where(roidb[0]['gt_classes'] != 0)[0] 42 | else: 43 | # For the COCO ground truth boxes, exclude the ones that are ''iscrowd'' 44 | gt_inds = np.where((roidb[0]['gt_classes'] != 0) & np.all(roidb[0]['gt_overlaps'].toarray() > -1.0, axis=1))[0] 45 | gt_boxes = np.empty((len(gt_inds), 5), dtype=np.float32) 46 | gt_boxes[:, 0:4] = roidb[0]['boxes'][gt_inds, :] * im_scales[0] 47 | gt_boxes[:, 4] = roidb[0]['gt_classes'][gt_inds] 48 | blobs['gt_boxes'] = gt_boxes 49 | blobs['im_info'] = np.array( 50 | [[im_blob.shape[1], im_blob.shape[2], im_scales[0]]], 51 | dtype=np.float32) 52 | 53 | handinfo = np.empty((len(gt_inds), 5), dtype=np.float32) 54 | handinfo[:,0] = roidb[0]['contactstate'][gt_inds] 55 | handinfo[:,1] = roidb[0]['handside'][gt_inds] 56 | handinfo[:,2] = roidb[0]['magnitude'][gt_inds] 57 | handinfo[:,3] = roidb[0]['unitdx'][gt_inds] 58 | handinfo[:,4] = roidb[0]['unitdy'][gt_inds] 59 | 60 | blobs['img_id'] = roidb[0]['img_id'] 61 | blobs['box_info'] = handinfo 62 | 63 | return blobs 64 | 65 | def _get_image_blob(roidb, scale_inds): 66 | """Builds an input blob from the images in the roidb at the specified 67 | scales. 68 | """ 69 | num_images = len(roidb) 70 | 71 | processed_ims = [] 72 | im_scales = [] 73 | for i in range(num_images): 74 | im = cv2.imread(roidb[i]['image']) 75 | # im = imread(roidb[i]['image']) 76 | 77 | # if len(im.shape) == 2: 78 | # im = im[:,:,np.newaxis] 79 | # im = np.concatenate((im,im,im), axis=2) 80 | # flip the channel, since the original one using cv2 81 | # rgb -> bgr 82 | # im = im[:,:,::-1] 83 | 84 | 85 | if roidb[i]['flipped']: 86 | im = im[:, ::-1, :] 87 | target_size = cfg.TRAIN.SCALES[scale_inds[i]] 88 | im, im_scale = prep_im_for_blob(im, cfg.PIXEL_MEANS, target_size, 89 | cfg.TRAIN.MAX_SIZE) 90 | im_scales.append(im_scale) 91 | processed_ims.append(im) 92 | 93 | # Create a blob to hold the input images 94 | blob = im_list_to_blob(processed_ims) 95 | 96 | return blob, im_scales 97 | -------------------------------------------------------------------------------- /lib/model/rpn/generate_anchors.py: -------------------------------------------------------------------------------- 1 | from __future__ import print_function 2 | # -------------------------------------------------------- 3 | # Faster R-CNN 4 | # Copyright (c) 2015 Microsoft 5 | # Licensed under The MIT License [see LICENSE for details] 6 | # Written by Ross Girshick and Sean Bell 7 | # -------------------------------------------------------- 8 | 9 | import numpy as np 10 | import pdb 11 | 12 | # Verify that we compute the same anchors as Shaoqing's matlab implementation: 13 | # 14 | # >> load output/rpn_cachedir/faster_rcnn_VOC2007_ZF_stage1_rpn/anchors.mat 15 | # >> anchors 16 | # 17 | # anchors = 18 | # 19 | # -83 -39 100 56 20 | # -175 -87 192 104 21 | # -359 -183 376 200 22 | # -55 -55 72 72 23 | # -119 -119 136 136 24 | # -247 -247 264 264 25 | # -35 -79 52 96 26 | # -79 -167 96 184 27 | # -167 -343 184 360 28 | 29 | #array([[ -83., -39., 100., 56.], 30 | # [-175., -87., 192., 104.], 31 | # [-359., -183., 376., 200.], 32 | # [ -55., -55., 72., 72.], 33 | # [-119., -119., 136., 136.], 34 | # [-247., -247., 264., 264.], 35 | # [ -35., -79., 52., 96.], 36 | # [ -79., -167., 96., 184.], 37 | # [-167., -343., 184., 360.]]) 38 | 39 | try: 40 | xrange # Python 2 41 | except NameError: 42 | xrange = range # Python 3 43 | 44 | 45 | def generate_anchors(base_size=16, ratios=[0.5, 1, 2], 46 | scales=2**np.arange(3, 6)): 47 | """ 48 | Generate anchor (reference) windows by enumerating aspect ratios X 49 | scales wrt a reference (0, 0, 15, 15) window. 50 | """ 51 | 52 | base_anchor = np.array([1, 1, base_size, base_size]) - 1 53 | ratio_anchors = _ratio_enum(base_anchor, ratios) 54 | anchors = np.vstack([_scale_enum(ratio_anchors[i, :], scales) 55 | for i in xrange(ratio_anchors.shape[0])]) 56 | return anchors 57 | 58 | def _whctrs(anchor): 59 | """ 60 | Return width, height, x center, and y center for an anchor (window). 61 | """ 62 | 63 | w = anchor[2] - anchor[0] + 1 64 | h = anchor[3] - anchor[1] + 1 65 | x_ctr = anchor[0] + 0.5 * (w - 1) 66 | y_ctr = anchor[1] + 0.5 * (h - 1) 67 | return w, h, x_ctr, y_ctr 68 | 69 | def _mkanchors(ws, hs, x_ctr, y_ctr): 70 | """ 71 | Given a vector of widths (ws) and heights (hs) around a center 72 | (x_ctr, y_ctr), output a set of anchors (windows). 73 | """ 74 | 75 | ws = ws[:, np.newaxis] 76 | hs = hs[:, np.newaxis] 77 | anchors = np.hstack((x_ctr - 0.5 * (ws - 1), 78 | y_ctr - 0.5 * (hs - 1), 79 | x_ctr + 0.5 * (ws - 1), 80 | y_ctr + 0.5 * (hs - 1))) 81 | return anchors 82 | 83 | def _ratio_enum(anchor, ratios): 84 | """ 85 | Enumerate a set of anchors for each aspect ratio wrt an anchor. 86 | """ 87 | 88 | w, h, x_ctr, y_ctr = _whctrs(anchor) 89 | size = w * h 90 | size_ratios = size / ratios 91 | ws = np.round(np.sqrt(size_ratios)) 92 | hs = np.round(ws * ratios) 93 | anchors = _mkanchors(ws, hs, x_ctr, y_ctr) 94 | return anchors 95 | 96 | def _scale_enum(anchor, scales): 97 | """ 98 | Enumerate a set of anchors for each scale wrt an anchor. 99 | """ 100 | 101 | w, h, x_ctr, y_ctr = _whctrs(anchor) 102 | ws = w * scales 103 | hs = h * scales 104 | anchors = _mkanchors(ws, hs, x_ctr, y_ctr) 105 | return anchors 106 | 107 | if __name__ == '__main__': 108 | import time 109 | t = time.time() 110 | a = generate_anchors() 111 | print(time.time() - t) 112 | print(a) 113 | from IPython import embed; embed() 114 | -------------------------------------------------------------------------------- /lib/model/utils/bbox.pyx: -------------------------------------------------------------------------------- 1 | # -------------------------------------------------------- 2 | # Fast R-CNN 3 | # Copyright (c) 2015 Microsoft 4 | # Licensed under The MIT License [see LICENSE for details] 5 | # Written by Sergey Karayev 6 | # -------------------------------------------------------- 7 | 8 | cimport cython 9 | import numpy as np 10 | cimport numpy as np 11 | 12 | DTYPE = np.float 13 | ctypedef np.float_t DTYPE_t 14 | 15 | def bbox_overlaps(np.ndarray[DTYPE_t, ndim=2] boxes, 16 | np.ndarray[DTYPE_t, ndim=2] query_boxes): 17 | return bbox_overlaps_c(boxes, query_boxes) 18 | 19 | cdef np.ndarray[DTYPE_t, ndim=2] bbox_overlaps_c( 20 | np.ndarray[DTYPE_t, ndim=2] boxes, 21 | np.ndarray[DTYPE_t, ndim=2] query_boxes): 22 | """ 23 | Parameters 24 | ---------- 25 | boxes: (N, 4) ndarray of float 26 | query_boxes: (K, 4) ndarray of float 27 | Returns 28 | ------- 29 | overlaps: (N, K) ndarray of overlap between boxes and query_boxes 30 | """ 31 | cdef unsigned int N = boxes.shape[0] 32 | cdef unsigned int K = query_boxes.shape[0] 33 | cdef np.ndarray[DTYPE_t, ndim=2] overlaps = np.zeros((N, K), dtype=DTYPE) 34 | cdef DTYPE_t iw, ih, box_area 35 | cdef DTYPE_t ua 36 | cdef unsigned int k, n 37 | for k in range(K): 38 | box_area = ( 39 | (query_boxes[k, 2] - query_boxes[k, 0] + 1) * 40 | (query_boxes[k, 3] - query_boxes[k, 1] + 1) 41 | ) 42 | for n in range(N): 43 | iw = ( 44 | min(boxes[n, 2], query_boxes[k, 2]) - 45 | max(boxes[n, 0], query_boxes[k, 0]) + 1 46 | ) 47 | if iw > 0: 48 | ih = ( 49 | min(boxes[n, 3], query_boxes[k, 3]) - 50 | max(boxes[n, 1], query_boxes[k, 1]) + 1 51 | ) 52 | if ih > 0: 53 | ua = float( 54 | (boxes[n, 2] - boxes[n, 0] + 1) * 55 | (boxes[n, 3] - boxes[n, 1] + 1) + 56 | box_area - iw * ih 57 | ) 58 | overlaps[n, k] = iw * ih / ua 59 | return overlaps 60 | 61 | 62 | def bbox_intersections( 63 | np.ndarray[DTYPE_t, ndim=2] boxes, 64 | np.ndarray[DTYPE_t, ndim=2] query_boxes): 65 | return bbox_intersections_c(boxes, query_boxes) 66 | 67 | 68 | cdef np.ndarray[DTYPE_t, ndim=2] bbox_intersections_c( 69 | np.ndarray[DTYPE_t, ndim=2] boxes, 70 | np.ndarray[DTYPE_t, ndim=2] query_boxes): 71 | """ 72 | For each query box compute the intersection ratio covered by boxes 73 | ---------- 74 | Parameters 75 | ---------- 76 | boxes: (N, 4) ndarray of float 77 | query_boxes: (K, 4) ndarray of float 78 | Returns 79 | ------- 80 | overlaps: (N, K) ndarray of intersec between boxes and query_boxes 81 | """ 82 | cdef unsigned int N = boxes.shape[0] 83 | cdef unsigned int K = query_boxes.shape[0] 84 | cdef np.ndarray[DTYPE_t, ndim=2] intersec = np.zeros((N, K), dtype=DTYPE) 85 | cdef DTYPE_t iw, ih, box_area 86 | cdef DTYPE_t ua 87 | cdef unsigned int k, n 88 | for k in range(K): 89 | box_area = ( 90 | (query_boxes[k, 2] - query_boxes[k, 0] + 1) * 91 | (query_boxes[k, 3] - query_boxes[k, 1] + 1) 92 | ) 93 | for n in range(N): 94 | iw = ( 95 | min(boxes[n, 2], query_boxes[k, 2]) - 96 | max(boxes[n, 0], query_boxes[k, 0]) + 1 97 | ) 98 | if iw > 0: 99 | ih = ( 100 | min(boxes[n, 3], query_boxes[k, 3]) - 101 | max(boxes[n, 1], query_boxes[k, 1]) + 1 102 | ) 103 | if ih > 0: 104 | intersec[n, k] = iw * ih / box_area 105 | return intersec -------------------------------------------------------------------------------- /lib/pycocotools/mask.py: -------------------------------------------------------------------------------- 1 | __author__ = 'tsungyi' 2 | 3 | from . import _mask 4 | 5 | # Interface for manipulating masks stored in RLE format. 6 | # 7 | # RLE is a simple yet efficient format for storing binary masks. RLE 8 | # first divides a vector (or vectorized image) into a series of piecewise 9 | # constant regions and then for each piece simply stores the length of 10 | # that piece. For example, given M=[0 0 1 1 1 0 1] the RLE counts would 11 | # be [2 3 1 1], or for M=[1 1 1 1 1 1 0] the counts would be [0 6 1] 12 | # (note that the odd counts are always the numbers of zeros). Instead of 13 | # storing the counts directly, additional compression is achieved with a 14 | # variable bitrate representation based on a common scheme called LEB128. 15 | # 16 | # Compression is greatest given large piecewise constant regions. 17 | # Specifically, the size of the RLE is proportional to the number of 18 | # *boundaries* in M (or for an image the number of boundaries in the y 19 | # direction). Assuming fairly simple shapes, the RLE representation is 20 | # O(sqrt(n)) where n is number of pixels in the object. Hence space usage 21 | # is substantially lower, especially for large simple objects (large n). 22 | # 23 | # Many common operations on masks can be computed directly using the RLE 24 | # (without need for decoding). This includes computations such as area, 25 | # union, intersection, etc. All of these operations are linear in the 26 | # size of the RLE, in other words they are O(sqrt(n)) where n is the area 27 | # of the object. Computing these operations on the original mask is O(n). 28 | # Thus, using the RLE can result in substantial computational savings. 29 | # 30 | # The following API functions are defined: 31 | # encode - Encode binary masks using RLE. 32 | # decode - Decode binary masks encoded via RLE. 33 | # merge - Compute union or intersection of encoded masks. 34 | # iou - Compute intersection over union between masks. 35 | # area - Compute area of encoded masks. 36 | # toBbox - Get bounding boxes surrounding encoded masks. 37 | # frPyObjects - Convert polygon, bbox, and uncompressed RLE to encoded RLE mask. 38 | # 39 | # Usage: 40 | # Rs = encode( masks ) 41 | # masks = decode( Rs ) 42 | # R = merge( Rs, intersect=false ) 43 | # o = iou( dt, gt, iscrowd ) 44 | # a = area( Rs ) 45 | # bbs = toBbox( Rs ) 46 | # Rs = frPyObjects( [pyObjects], h, w ) 47 | # 48 | # In the API the following formats are used: 49 | # Rs - [dict] Run-length encoding of binary masks 50 | # R - dict Run-length encoding of binary mask 51 | # masks - [hxwxn] Binary mask(s) (must have type np.ndarray(dtype=uint8) in column-major order) 52 | # iscrowd - [nx1] list of np.ndarray. 1 indicates corresponding gt image has crowd region to ignore 53 | # bbs - [nx4] Bounding box(es) stored as [x y w h] 54 | # poly - Polygon stored as [[x1 y1 x2 y2...],[x1 y1 ...],...] (2D list) 55 | # dt,gt - May be either bounding boxes or encoded masks 56 | # Both poly and bbs are 0-indexed (bbox=[0 0 1 1] encloses first pixel). 57 | # 58 | # Finally, a note about the intersection over union (iou) computation. 59 | # The standard iou of a ground truth (gt) and detected (dt) object is 60 | # iou(gt,dt) = area(intersect(gt,dt)) / area(union(gt,dt)) 61 | # For "crowd" regions, we use a modified criteria. If a gt object is 62 | # marked as "iscrowd", we allow a dt to match any subregion of the gt. 63 | # Choosing gt' in the crowd gt that best matches the dt can be done using 64 | # gt'=intersect(dt,gt). Since by definition union(gt',dt)=dt, computing 65 | # iou(gt,dt,iscrowd) = iou(gt',dt) = area(intersect(gt,dt)) / area(dt) 66 | # For crowd gt regions we use this modified criteria above for the iou. 67 | # 68 | # To compile run "python setup.py build_ext --inplace" 69 | # Please do not contact us for help with compiling. 70 | # 71 | # Microsoft COCO Toolbox. version 2.0 72 | # Data, paper, and tutorials available at: http://mscoco.org/ 73 | # Code written by Piotr Dollar and Tsung-Yi Lin, 2015. 74 | # Licensed under the Simplified BSD License [see coco/license.txt] 75 | 76 | encode = _mask.encode 77 | decode = _mask.decode 78 | iou = _mask.iou 79 | merge = _mask.merge 80 | area = _mask.area 81 | toBbox = _mask.toBbox 82 | frPyObjects = _mask.frPyObjects -------------------------------------------------------------------------------- /.gitignore: -------------------------------------------------------------------------------- 1 | # READ THIS BEFORE YOU REFACTOR ME 2 | # 3 | # setup.py uses the list of patterns in this file to decide 4 | # what to delete, but it's not 100% sound. So, for example, 5 | # if you delete aten/build/ because it's redundant with build/, 6 | # aten/build/ will stop being cleaned. So be careful when 7 | # refactoring this file! 8 | 9 | ## PyTorch 10 | 11 | data/ 12 | data_preparation/ 13 | data_preparation_no_objectionable/ 14 | data_to_release/ 15 | demo/ 16 | compare/ 17 | 18 | images/ 19 | images_det/ 20 | 21 | models/* 22 | predata_dandan/multitask/*.json 23 | predata_dandan/ 24 | 25 | 26 | output/ 27 | lib/datasets/compare_eval/ 28 | backup/ 29 | pick_2/ 30 | images/ 31 | logs/ 32 | 33 | README_details.md 34 | debug.py 35 | demo_video.py 36 | verify_my_code.py 37 | 38 | .mypy_cache 39 | *.pyc 40 | */*.pyc 41 | */*.so* 42 | */**/__pycache__ 43 | */**/*.dylib* 44 | */**/*.pyc 45 | */**/*.pyd 46 | */**/*.so* 47 | */**/**/*.pyc 48 | */**/**/**/*.pyc 49 | */**/**/**/**/*.pyc 50 | aten/build/ 51 | aten/src/ATen/Config.h 52 | aten/src/ATen/cuda/CUDAConfig.h 53 | build/ 54 | dist/ 55 | docs/src/**/* 56 | test/.coverage 57 | test/cpp/api/mnist 58 | test/data/gpu_tensors.pt 59 | test/data/legacy_modules.t7 60 | test/data/legacy_serialized.pt 61 | test/data/linear.pt 62 | test/htmlcov 63 | third_party/build/ 64 | tools/shared/_utils_internal.py 65 | torch.egg-info/ 66 | torch/csrc/autograd/generated/* 67 | torch/csrc/cudnn/cuDNN.cpp 68 | torch/csrc/generated 69 | torch/csrc/generic/TensorMethods.cpp 70 | torch/csrc/jit/generated/* 71 | torch/csrc/nn/THCUNN.cpp 72 | torch/csrc/nn/THCUNN.cwrap 73 | torch/csrc/nn/THNN_generic.cpp 74 | torch/csrc/nn/THNN_generic.cwrap 75 | torch/csrc/nn/THNN_generic.h 76 | torch/csrc/nn/THNN.cpp 77 | torch/csrc/nn/THNN.cwrap 78 | torch/lib/*.a* 79 | torch/lib/*.dll* 80 | torch/lib/*.dylib* 81 | torch/lib/*.h 82 | torch/lib/*.lib 83 | torch/lib/*.so* 84 | torch/lib/build 85 | torch/lib/cmake 86 | torch/lib/include 87 | torch/lib/pkgconfig 88 | torch/lib/protoc 89 | torch/lib/tmp_install 90 | torch/lib/torch_shm_manager 91 | torch/version.py 92 | 93 | # IPython notebook checkpoints 94 | .ipynb_checkpoints 95 | 96 | # Editor temporaries 97 | *.swn 98 | *.swo 99 | *.swp 100 | *.swm 101 | *~ 102 | 103 | # macOS dir files 104 | .DS_Store 105 | 106 | # Symbolic files 107 | tools/shared/cwrap_common.py 108 | 109 | # Ninja files 110 | .ninja_deps 111 | .ninja_log 112 | compile_commands.json 113 | *.egg-info/ 114 | docs/source/scripts/activation_images/ 115 | 116 | ## General 117 | 118 | # Compiled Object files 119 | *.slo 120 | *.lo 121 | *.o 122 | *.cuo 123 | *.obj 124 | 125 | # Compiled Dynamic libraries 126 | *.so 127 | *.dylib 128 | *.dll 129 | 130 | # Compiled Static libraries 131 | *.lai 132 | *.la 133 | *.a 134 | *.lib 135 | 136 | # Compiled protocol buffers 137 | *.pb.h 138 | *.pb.cc 139 | *_pb2.py 140 | 141 | # Compiled python 142 | *.pyc 143 | *.pyd 144 | 145 | # Compiled MATLAB 146 | *.mex* 147 | 148 | # IPython notebook checkpoints 149 | .ipynb_checkpoints 150 | 151 | # Editor temporaries 152 | *.swn 153 | *.swo 154 | *.swp 155 | *~ 156 | 157 | # Sublime Text settings 158 | *.sublime-workspace 159 | *.sublime-project 160 | 161 | # Eclipse Project settings 162 | *.*project 163 | .settings 164 | 165 | # QtCreator files 166 | *.user 167 | 168 | # PyCharm files 169 | .idea 170 | 171 | # Visual Studio Code files 172 | .vscode 173 | .vs 174 | 175 | # OSX dir files 176 | .DS_Store 177 | 178 | ## Caffe2 179 | 180 | # build, distribute, and bins (+ python proto bindings) 181 | build 182 | build_host_protoc 183 | build_android 184 | build_ios 185 | /build_* 186 | .build_debug/* 187 | .build_release/* 188 | distribute/* 189 | *.testbin 190 | *.bin 191 | cmake_build 192 | .cmake_build 193 | gen 194 | .setuptools-cmake-build 195 | .pytest_cache 196 | aten/build/* 197 | 198 | # Bram 199 | plsdontbreak 200 | 201 | # Generated documentation 202 | docs/_site 203 | docs/gathered 204 | _site 205 | doxygen 206 | docs/dev 207 | 208 | # LevelDB files 209 | *.sst 210 | *.ldb 211 | LOCK 212 | LOG* 213 | CURRENT 214 | MANIFEST-* 215 | 216 | # generated version file 217 | caffe2/version.py 218 | 219 | # setup.py intermediates 220 | .eggs 221 | caffe2.egg-info 222 | 223 | # Atom/Watchman required file 224 | .watchmanconfig 225 | 226 | # cython generated files 227 | lib/model/utils/bbox.c 228 | lib/pycocotools/_mask.c -------------------------------------------------------------------------------- /lib/model/roi_pooling/src/roi_pooling.c: -------------------------------------------------------------------------------- 1 | #include 2 | #include 3 | 4 | int roi_pooling_forward(int pooled_height, int pooled_width, float spatial_scale, 5 | THFloatTensor * features, THFloatTensor * rois, THFloatTensor * output) 6 | { 7 | // Grab the input tensor 8 | float * data_flat = THFloatTensor_data(features); 9 | float * rois_flat = THFloatTensor_data(rois); 10 | 11 | float * output_flat = THFloatTensor_data(output); 12 | 13 | // Number of ROIs 14 | int num_rois = THFloatTensor_size(rois, 0); 15 | int size_rois = THFloatTensor_size(rois, 1); 16 | // batch size 17 | int batch_size = THFloatTensor_size(features, 0); 18 | if(batch_size != 1) 19 | { 20 | return 0; 21 | } 22 | // data height 23 | int data_height = THFloatTensor_size(features, 1); 24 | // data width 25 | int data_width = THFloatTensor_size(features, 2); 26 | // Number of channels 27 | int num_channels = THFloatTensor_size(features, 3); 28 | 29 | // Set all element of the output tensor to -inf. 30 | THFloatStorage_fill(THFloatTensor_storage(output), -1); 31 | 32 | // For each ROI R = [batch_index x1 y1 x2 y2]: max pool over R 33 | int index_roi = 0; 34 | int index_output = 0; 35 | int n; 36 | for (n = 0; n < num_rois; ++n) 37 | { 38 | int roi_batch_ind = rois_flat[index_roi + 0]; 39 | int roi_start_w = round(rois_flat[index_roi + 1] * spatial_scale); 40 | int roi_start_h = round(rois_flat[index_roi + 2] * spatial_scale); 41 | int roi_end_w = round(rois_flat[index_roi + 3] * spatial_scale); 42 | int roi_end_h = round(rois_flat[index_roi + 4] * spatial_scale); 43 | // CHECK_GE(roi_batch_ind, 0); 44 | // CHECK_LT(roi_batch_ind, batch_size); 45 | 46 | int roi_height = fmaxf(roi_end_h - roi_start_h + 1, 1); 47 | int roi_width = fmaxf(roi_end_w - roi_start_w + 1, 1); 48 | float bin_size_h = (float)(roi_height) / (float)(pooled_height); 49 | float bin_size_w = (float)(roi_width) / (float)(pooled_width); 50 | 51 | int index_data = roi_batch_ind * data_height * data_width * num_channels; 52 | const int output_area = pooled_width * pooled_height; 53 | 54 | int c, ph, pw; 55 | for (ph = 0; ph < pooled_height; ++ph) 56 | { 57 | for (pw = 0; pw < pooled_width; ++pw) 58 | { 59 | int hstart = (floor((float)(ph) * bin_size_h)); 60 | int wstart = (floor((float)(pw) * bin_size_w)); 61 | int hend = (ceil((float)(ph + 1) * bin_size_h)); 62 | int wend = (ceil((float)(pw + 1) * bin_size_w)); 63 | 64 | hstart = fminf(fmaxf(hstart + roi_start_h, 0), data_height); 65 | hend = fminf(fmaxf(hend + roi_start_h, 0), data_height); 66 | wstart = fminf(fmaxf(wstart + roi_start_w, 0), data_width); 67 | wend = fminf(fmaxf(wend + roi_start_w, 0), data_width); 68 | 69 | const int pool_index = index_output + (ph * pooled_width + pw); 70 | int is_empty = (hend <= hstart) || (wend <= wstart); 71 | if (is_empty) 72 | { 73 | for (c = 0; c < num_channels * output_area; c += output_area) 74 | { 75 | output_flat[pool_index + c] = 0; 76 | } 77 | } 78 | else 79 | { 80 | int h, w, c; 81 | for (h = hstart; h < hend; ++h) 82 | { 83 | for (w = wstart; w < wend; ++w) 84 | { 85 | for (c = 0; c < num_channels; ++c) 86 | { 87 | const int index = (h * data_width + w) * num_channels + c; 88 | if (data_flat[index_data + index] > output_flat[pool_index + c * output_area]) 89 | { 90 | output_flat[pool_index + c * output_area] = data_flat[index_data + index]; 91 | } 92 | } 93 | } 94 | } 95 | } 96 | } 97 | } 98 | 99 | // Increment ROI index 100 | index_roi += size_rois; 101 | index_output += pooled_height * pooled_width * num_channels; 102 | } 103 | return 1; 104 | } -------------------------------------------------------------------------------- /lib/datasets/vg_eval.py: -------------------------------------------------------------------------------- 1 | from __future__ import absolute_import 2 | # -------------------------------------------------------- 3 | # Fast/er R-CNN 4 | # Licensed under The MIT License [see LICENSE for details] 5 | # Written by Bharath Hariharan 6 | # -------------------------------------------------------- 7 | 8 | import xml.etree.ElementTree as ET 9 | import os 10 | import numpy as np 11 | from .voc_eval import voc_ap 12 | 13 | def vg_eval( detpath, 14 | gt_roidb, 15 | image_index, 16 | classindex, 17 | ovthresh=0.5, 18 | use_07_metric=False, 19 | eval_attributes=False): 20 | """rec, prec, ap, sorted_scores, npos = voc_eval( 21 | detpath, 22 | gt_roidb, 23 | image_index, 24 | classindex, 25 | [ovthresh], 26 | [use_07_metric]) 27 | 28 | Top level function that does the Visual Genome evaluation. 29 | 30 | detpath: Path to detections 31 | gt_roidb: List of ground truth structs. 32 | image_index: List of image ids. 33 | classindex: Category index 34 | [ovthresh]: Overlap threshold (default = 0.5) 35 | [use_07_metric]: Whether to use VOC07's 11 point AP computation 36 | (default False) 37 | """ 38 | # extract gt objects for this class 39 | class_recs = {} 40 | npos = 0 41 | for item,imagename in zip(gt_roidb,image_index): 42 | if eval_attributes: 43 | bbox = item['boxes'][np.where(np.any(item['gt_attributes'].toarray() == classindex, axis=1))[0], :] 44 | else: 45 | bbox = item['boxes'][np.where(item['gt_classes'] == classindex)[0], :] 46 | difficult = np.zeros((bbox.shape[0],)).astype(np.bool) 47 | det = [False] * bbox.shape[0] 48 | npos = npos + sum(~difficult) 49 | class_recs[str(imagename)] = {'bbox': bbox, 50 | 'difficult': difficult, 51 | 'det': det} 52 | if npos == 0: 53 | # No ground truth examples 54 | return 0,0,0,0,npos 55 | 56 | # read dets 57 | with open(detpath, 'r') as f: 58 | lines = f.readlines() 59 | if len(lines) == 0: 60 | # No detection examples 61 | return 0,0,0,0,npos 62 | 63 | splitlines = [x.strip().split(' ') for x in lines] 64 | image_ids = [x[0] for x in splitlines] 65 | confidence = np.array([float(x[1]) for x in splitlines]) 66 | BB = np.array([[float(z) for z in x[2:]] for x in splitlines]) 67 | 68 | # sort by confidence 69 | sorted_ind = np.argsort(-confidence) 70 | sorted_scores = -np.sort(-confidence) 71 | BB = BB[sorted_ind, :] 72 | image_ids = [image_ids[x] for x in sorted_ind] 73 | 74 | # go down dets and mark TPs and FPs 75 | nd = len(image_ids) 76 | tp = np.zeros(nd) 77 | fp = np.zeros(nd) 78 | for d in range(nd): 79 | R = class_recs[image_ids[d]] 80 | bb = BB[d, :].astype(float) 81 | ovmax = -np.inf 82 | BBGT = R['bbox'].astype(float) 83 | 84 | if BBGT.size > 0: 85 | # compute overlaps 86 | # intersection 87 | ixmin = np.maximum(BBGT[:, 0], bb[0]) 88 | iymin = np.maximum(BBGT[:, 1], bb[1]) 89 | ixmax = np.minimum(BBGT[:, 2], bb[2]) 90 | iymax = np.minimum(BBGT[:, 3], bb[3]) 91 | iw = np.maximum(ixmax - ixmin + 1., 0.) 92 | ih = np.maximum(iymax - iymin + 1., 0.) 93 | inters = iw * ih 94 | 95 | # union 96 | uni = ((bb[2] - bb[0] + 1.) * (bb[3] - bb[1] + 1.) + 97 | (BBGT[:, 2] - BBGT[:, 0] + 1.) * 98 | (BBGT[:, 3] - BBGT[:, 1] + 1.) - inters) 99 | 100 | overlaps = inters / uni 101 | ovmax = np.max(overlaps) 102 | jmax = np.argmax(overlaps) 103 | 104 | if ovmax > ovthresh: 105 | if not R['difficult'][jmax]: 106 | if not R['det'][jmax]: 107 | tp[d] = 1. 108 | R['det'][jmax] = 1 109 | else: 110 | fp[d] = 1. 111 | else: 112 | fp[d] = 1. 113 | 114 | # compute precision recall 115 | fp = np.cumsum(fp) 116 | tp = np.cumsum(tp) 117 | rec = tp / float(npos) 118 | # avoid divide by zero in case the first detection matches a difficult 119 | # ground truth 120 | prec = tp / np.maximum(tp + fp, np.finfo(np.float64).eps) 121 | ap = voc_ap(rec, prec, use_07_metric) 122 | 123 | return rec, prec, ap, sorted_scores, npos 124 | -------------------------------------------------------------------------------- /lib/model/rpn/rpn.py: -------------------------------------------------------------------------------- 1 | from __future__ import absolute_import 2 | import torch 3 | import torch.nn as nn 4 | import torch.nn.functional as F 5 | from torch.autograd import Variable 6 | 7 | from model.utils.config import cfg 8 | from .proposal_layer import _ProposalLayer 9 | from .anchor_target_layer import _AnchorTargetLayer 10 | from model.utils.net_utils import _smooth_l1_loss 11 | 12 | import numpy as np 13 | import math 14 | import pdb 15 | import time 16 | 17 | class _RPN(nn.Module): 18 | """ region proposal network """ 19 | def __init__(self, din): 20 | super(_RPN, self).__init__() 21 | 22 | self.din = din # get depth of input feature map, e.g., 512 23 | self.anchor_scales = cfg.ANCHOR_SCALES 24 | self.anchor_ratios = cfg.ANCHOR_RATIOS 25 | self.feat_stride = cfg.FEAT_STRIDE[0] 26 | 27 | # define the convrelu layers processing input feature map 28 | self.RPN_Conv = nn.Conv2d(self.din, 512, 3, 1, 1, bias=True) 29 | 30 | # define bg/fg classifcation score layer 31 | self.nc_score_out = len(self.anchor_scales) * len(self.anchor_ratios) * 2 # 2(bg/fg) * 9 (anchors) 32 | self.RPN_cls_score = nn.Conv2d(512, self.nc_score_out, 1, 1, 0) 33 | 34 | # define anchor box offset prediction layer 35 | self.nc_bbox_out = len(self.anchor_scales) * len(self.anchor_ratios) * 4 # 4(coords) * 9 (anchors) 36 | self.RPN_bbox_pred = nn.Conv2d(512, self.nc_bbox_out, 1, 1, 0) 37 | 38 | # define proposal layer 39 | self.RPN_proposal = _ProposalLayer(self.feat_stride, self.anchor_scales, self.anchor_ratios) 40 | 41 | # define anchor target layer 42 | self.RPN_anchor_target = _AnchorTargetLayer(self.feat_stride, self.anchor_scales, self.anchor_ratios) 43 | 44 | self.rpn_loss_cls = 0 45 | self.rpn_loss_box = 0 46 | 47 | @staticmethod 48 | def reshape(x, d): 49 | input_shape = x.size() 50 | x = x.view( 51 | input_shape[0], 52 | int(d), 53 | int(float(input_shape[1] * input_shape[2]) / float(d)), 54 | input_shape[3] 55 | ) 56 | return x 57 | 58 | def forward(self, base_feat, im_info, gt_boxes, num_boxes): 59 | 60 | batch_size = base_feat.size(0) 61 | 62 | # return feature map after convrelu layer 63 | rpn_conv1 = F.relu(self.RPN_Conv(base_feat), inplace=True) 64 | # get rpn classification score 65 | rpn_cls_score = self.RPN_cls_score(rpn_conv1) 66 | 67 | rpn_cls_score_reshape = self.reshape(rpn_cls_score, 2) 68 | rpn_cls_prob_reshape = F.softmax(rpn_cls_score_reshape, 1) 69 | rpn_cls_prob = self.reshape(rpn_cls_prob_reshape, self.nc_score_out) 70 | 71 | # get rpn offsets to the anchor boxes 72 | rpn_bbox_pred = self.RPN_bbox_pred(rpn_conv1) 73 | 74 | # proposal layer 75 | cfg_key = 'TRAIN' if self.training else 'TEST' 76 | 77 | rois = self.RPN_proposal((rpn_cls_prob.data, rpn_bbox_pred.data, 78 | im_info, cfg_key)) 79 | 80 | self.rpn_loss_cls = 0 81 | self.rpn_loss_box = 0 82 | 83 | # generating training labels and build the rpn loss 84 | if self.training: 85 | assert gt_boxes is not None 86 | 87 | rpn_data = self.RPN_anchor_target((rpn_cls_score.data, gt_boxes, im_info, num_boxes)) 88 | 89 | # compute classification loss 90 | rpn_cls_score = rpn_cls_score_reshape.permute(0, 2, 3, 1).contiguous().view(batch_size, -1, 2) 91 | rpn_label = rpn_data[0].view(batch_size, -1) 92 | 93 | rpn_keep = Variable(rpn_label.view(-1).ne(-1).nonzero().view(-1)) 94 | rpn_cls_score = torch.index_select(rpn_cls_score.view(-1,2), 0, rpn_keep) 95 | rpn_label = torch.index_select(rpn_label.view(-1), 0, rpn_keep.data) 96 | rpn_label = Variable(rpn_label.long()) 97 | self.rpn_loss_cls = F.cross_entropy(rpn_cls_score, rpn_label) 98 | fg_cnt = torch.sum(rpn_label.data.ne(0)) 99 | 100 | rpn_bbox_targets, rpn_bbox_inside_weights, rpn_bbox_outside_weights = rpn_data[1:] 101 | 102 | # compute bbox regression loss 103 | rpn_bbox_inside_weights = Variable(rpn_bbox_inside_weights) 104 | rpn_bbox_outside_weights = Variable(rpn_bbox_outside_weights) 105 | rpn_bbox_targets = Variable(rpn_bbox_targets) 106 | 107 | self.rpn_loss_box = _smooth_l1_loss(rpn_bbox_pred, rpn_bbox_targets, rpn_bbox_inside_weights, 108 | rpn_bbox_outside_weights, sigma=3, dim=[1,2,3]) 109 | 110 | return rois, self.rpn_loss_cls, self.rpn_loss_box 111 | -------------------------------------------------------------------------------- /lib/roi_data_layer/roidb.py: -------------------------------------------------------------------------------- 1 | """Transform a roidb into a trainable roidb by adding a bunch of metadata.""" 2 | from __future__ import absolute_import 3 | from __future__ import division 4 | from __future__ import print_function 5 | 6 | import datasets 7 | import numpy as np 8 | from model.utils.config import cfg 9 | from datasets.factory import get_imdb 10 | import PIL 11 | import pdb 12 | 13 | def prepare_roidb(imdb): 14 | """Enrich the imdb's roidb by adding some derived quantities that 15 | are useful for training. This function precomputes the maximum 16 | overlap, taken over ground-truth boxes, between each ROI and 17 | each ground-truth box. The class with maximum overlap is also 18 | recorded. 19 | """ 20 | 21 | roidb = imdb.roidb 22 | if not (imdb.name.startswith('coco')): 23 | sizes = [PIL.Image.open(imdb.image_path_at(i)).size 24 | for i in range(imdb.num_images)] 25 | 26 | for i in range(len(imdb.image_index)): 27 | roidb[i]['img_id'] = imdb.image_id_at(i) 28 | roidb[i]['image'] = imdb.image_path_at(i) 29 | if not (imdb.name.startswith('coco')): 30 | roidb[i]['width'] = sizes[i][0] 31 | roidb[i]['height'] = sizes[i][1] 32 | # need gt_overlaps as a dense array for argmax 33 | gt_overlaps = roidb[i]['gt_overlaps'].toarray() 34 | # max overlap with gt over classes (columns) 35 | max_overlaps = gt_overlaps.max(axis=1) 36 | # gt class that had the max overlap 37 | max_classes = gt_overlaps.argmax(axis=1) 38 | roidb[i]['max_classes'] = max_classes 39 | roidb[i]['max_overlaps'] = max_overlaps 40 | # sanity checks 41 | # max overlap of 0 => class should be zero (background) 42 | zero_inds = np.where(max_overlaps == 0)[0] 43 | assert all(max_classes[zero_inds] == 0) 44 | # max overlap > 0 => class should not be zero (must be a fg class) 45 | nonzero_inds = np.where(max_overlaps > 0)[0] 46 | assert all(max_classes[nonzero_inds] != 0) 47 | 48 | 49 | def rank_roidb_ratio(roidb): 50 | # rank roidb based on the ratio between width and height. 51 | ratio_large = 2 # largest ratio to preserve. 52 | ratio_small = 0.5 # smallest ratio to preserve. 53 | 54 | ratio_list = [] 55 | for i in range(len(roidb)): 56 | width = roidb[i]['width'] 57 | height = roidb[i]['height'] 58 | ratio = width / float(height) 59 | 60 | if ratio > ratio_large: 61 | roidb[i]['need_crop'] = 1 62 | ratio = ratio_large 63 | elif ratio < ratio_small: 64 | roidb[i]['need_crop'] = 1 65 | ratio = ratio_small 66 | else: 67 | roidb[i]['need_crop'] = 0 68 | 69 | ratio_list.append(ratio) 70 | 71 | ratio_list = np.array(ratio_list) 72 | ratio_index = np.argsort(ratio_list) 73 | return ratio_list[ratio_index], ratio_index 74 | 75 | def filter_roidb(roidb): 76 | # filter the image without bounding box. 77 | print('before filtering, there are %d images...' % (len(roidb))) 78 | i = 0 79 | while i < len(roidb): 80 | if len(roidb[i]['boxes']) == 0: 81 | del roidb[i] 82 | i -= 1 83 | i += 1 84 | 85 | print('after filtering, there are %d images...' % (len(roidb))) 86 | return roidb 87 | 88 | def combined_roidb(imdb_names, training=True, leftright=False): 89 | """ 90 | Combine multiple roidbs 91 | """ 92 | 93 | def get_training_roidb(imdb, leftright=False): 94 | """Returns a roidb (Region of Interest database) for use in training.""" 95 | if cfg.TRAIN.USE_FLIPPED: 96 | if leftright: 97 | print('Appending horizontally-flipped training examples...') 98 | imdb.append_flipped_images() 99 | print('done') 100 | else: 101 | print('Appending horizontally-flipped training examples...') 102 | imdb.append_flipped_images(leftright) 103 | print('done') 104 | 105 | 106 | print('Preparing training data...') 107 | 108 | prepare_roidb(imdb) 109 | #ratio_index = rank_roidb_ratio(imdb) 110 | print('done') 111 | 112 | return imdb.roidb 113 | 114 | 115 | def get_roidb(imdb_name): 116 | imdb = get_imdb(imdb_name) 117 | print('Loaded dataset `{:s}` for training'.format(imdb.name)) 118 | imdb.set_proposal_method(cfg.TRAIN.PROPOSAL_METHOD) 119 | print('Set proposal method: {:s}'.format(cfg.TRAIN.PROPOSAL_METHOD)) 120 | print(imdb_name) 121 | roidb = get_training_roidb(imdb) 122 | return roidb 123 | 124 | roidbs = [get_roidb(s) for s in imdb_names.split('+')] 125 | roidb = roidbs[0] 126 | 127 | if len(roidbs) > 1: 128 | for r in roidbs[1:]: 129 | roidb.extend(r) 130 | tmp = get_imdb(imdb_names.split('+')[1]) 131 | imdb = datasets.imdb.imdb(imdb_names, tmp.classes) 132 | else: 133 | imdb = get_imdb(imdb_names) 134 | 135 | if training: 136 | roidb = filter_roidb(roidb) 137 | 138 | ratio_list, ratio_index = rank_roidb_ratio(roidb) 139 | 140 | return imdb, roidb, ratio_list, ratio_index 141 | -------------------------------------------------------------------------------- /lib/model/roi_crop/src/roi_crop_cuda.c: -------------------------------------------------------------------------------- 1 | #include 2 | #include 3 | #include 4 | #include "roi_crop_cuda_kernel.h" 5 | 6 | #define real float 7 | 8 | // this symbol will be resolved automatically from PyTorch libs 9 | extern THCState *state; 10 | 11 | // Bilinear sampling is done in BHWD (coalescing is not obvious in BDHW) 12 | // we assume BHWD format in inputImages 13 | // we assume BHW(YX) format on grids 14 | 15 | int BilinearSamplerBHWD_updateOutput_cuda(THCudaTensor *inputImages, THCudaTensor *grids, THCudaTensor *output){ 16 | // THCState *state = getCutorchState(L); 17 | // THCudaTensor *inputImages = (THCudaTensor *)luaT_checkudata(L, 2, "torch.CudaTensor"); 18 | // THCudaTensor *grids = (THCudaTensor *)luaT_checkudata(L, 3, "torch.CudaTensor"); 19 | // THCudaTensor *output = (THCudaTensor *)luaT_checkudata(L, 4, "torch.CudaTensor"); 20 | 21 | int success = 0; 22 | success = BilinearSamplerBHWD_updateOutput_cuda_kernel(THCudaTensor_size(state, output, 1), 23 | THCudaTensor_size(state, output, 3), 24 | THCudaTensor_size(state, output, 2), 25 | THCudaTensor_size(state, output, 0), 26 | THCudaTensor_size(state, inputImages, 1), 27 | THCudaTensor_size(state, inputImages, 2), 28 | THCudaTensor_size(state, inputImages, 3), 29 | THCudaTensor_size(state, inputImages, 0), 30 | THCudaTensor_data(state, inputImages), 31 | THCudaTensor_stride(state, inputImages, 0), 32 | THCudaTensor_stride(state, inputImages, 1), 33 | THCudaTensor_stride(state, inputImages, 2), 34 | THCudaTensor_stride(state, inputImages, 3), 35 | THCudaTensor_data(state, grids), 36 | THCudaTensor_stride(state, grids, 0), 37 | THCudaTensor_stride(state, grids, 3), 38 | THCudaTensor_stride(state, grids, 1), 39 | THCudaTensor_stride(state, grids, 2), 40 | THCudaTensor_data(state, output), 41 | THCudaTensor_stride(state, output, 0), 42 | THCudaTensor_stride(state, output, 1), 43 | THCudaTensor_stride(state, output, 2), 44 | THCudaTensor_stride(state, output, 3), 45 | THCState_getCurrentStream(state)); 46 | 47 | //check for errors 48 | if (!success) { 49 | THError("aborting"); 50 | } 51 | return 1; 52 | } 53 | 54 | int BilinearSamplerBHWD_updateGradInput_cuda(THCudaTensor *inputImages, THCudaTensor *grids, THCudaTensor *gradInputImages, 55 | THCudaTensor *gradGrids, THCudaTensor *gradOutput) 56 | { 57 | // THCState *state = getCutorchState(L); 58 | // THCudaTensor *inputImages = (THCudaTensor *)luaT_checkudata(L, 2, "torch.CudaTensor"); 59 | // THCudaTensor *grids = (THCudaTensor *)luaT_checkudata(L, 3, "torch.CudaTensor"); 60 | // THCudaTensor *gradInputImages = (THCudaTensor *)luaT_checkudata(L, 4, "torch.CudaTensor"); 61 | // THCudaTensor *gradGrids = (THCudaTensor *)luaT_checkudata(L, 5, "torch.CudaTensor"); 62 | // THCudaTensor *gradOutput = (THCudaTensor *)luaT_checkudata(L, 6, "torch.CudaTensor"); 63 | 64 | int success = 0; 65 | success = BilinearSamplerBHWD_updateGradInput_cuda_kernel(THCudaTensor_size(state, gradOutput, 1), 66 | THCudaTensor_size(state, gradOutput, 3), 67 | THCudaTensor_size(state, gradOutput, 2), 68 | THCudaTensor_size(state, gradOutput, 0), 69 | THCudaTensor_size(state, inputImages, 1), 70 | THCudaTensor_size(state, inputImages, 2), 71 | THCudaTensor_size(state, inputImages, 3), 72 | THCudaTensor_size(state, inputImages, 0), 73 | THCudaTensor_data(state, inputImages), 74 | THCudaTensor_stride(state, inputImages, 0), 75 | THCudaTensor_stride(state, inputImages, 1), 76 | THCudaTensor_stride(state, inputImages, 2), 77 | THCudaTensor_stride(state, inputImages, 3), 78 | THCudaTensor_data(state, grids), 79 | THCudaTensor_stride(state, grids, 0), 80 | THCudaTensor_stride(state, grids, 3), 81 | THCudaTensor_stride(state, grids, 1), 82 | THCudaTensor_stride(state, grids, 2), 83 | THCudaTensor_data(state, gradInputImages), 84 | THCudaTensor_stride(state, gradInputImages, 0), 85 | THCudaTensor_stride(state, gradInputImages, 1), 86 | THCudaTensor_stride(state, gradInputImages, 2), 87 | THCudaTensor_stride(state, gradInputImages, 3), 88 | THCudaTensor_data(state, gradGrids), 89 | THCudaTensor_stride(state, gradGrids, 0), 90 | THCudaTensor_stride(state, gradGrids, 3), 91 | THCudaTensor_stride(state, gradGrids, 1), 92 | THCudaTensor_stride(state, gradGrids, 2), 93 | THCudaTensor_data(state, gradOutput), 94 | THCudaTensor_stride(state, gradOutput, 0), 95 | THCudaTensor_stride(state, gradOutput, 1), 96 | THCudaTensor_stride(state, gradOutput, 2), 97 | THCudaTensor_stride(state, gradOutput, 3), 98 | THCState_getCurrentStream(state)); 99 | 100 | //check for errors 101 | if (!success) { 102 | THError("aborting"); 103 | } 104 | return 1; 105 | } 106 | -------------------------------------------------------------------------------- /lib/model/extension_layers/extension_layers.py: -------------------------------------------------------------------------------- 1 | import torch 2 | from torch import nn 3 | from torch.autograd import Function 4 | from torch.autograd.function import once_differentiable 5 | from torch.nn.modules.utils import _pair 6 | import torch.nn.functional as F 7 | import pickle 8 | import datetime 9 | 10 | 11 | class extension_layer(nn.Module): 12 | def __init__(self): 13 | super(extension_layer, self).__init__() 14 | self.init_layers_weights() 15 | 16 | 17 | def forward(self, input, input_padded, roi_labels, box_info): 18 | if (len(input.shape)) == 2: 19 | input = input.unsqueeze(0) 20 | 21 | if (len(input_padded.shape)) == 2: 22 | input_padded = input_padded.unsqueeze(0) 23 | 24 | loss_list = [self.hand_contactstate_part(input_padded, roi_labels, box_info), \ 25 | self.hand_dxdymagnitude_part(input_padded, roi_labels, box_info),\ 26 | self.hand_handside_part(input, roi_labels, box_info)] 27 | 28 | return loss_list 29 | 30 | def init_layers_weights(self): 31 | 32 | self.hand_contact_state_layer = nn.Sequential(nn.Linear(2048, 32), \ 33 | nn.ReLU(), \ 34 | nn.Dropout(p=0.5),\ 35 | nn.Linear(32, 5)) 36 | self.hand_dydx_layer = torch.nn.Linear(2048, 3) 37 | self.hand_lr_layer = torch.nn.Linear(2048, 1) 38 | 39 | 40 | self.hand_contactstate_loss = nn.CrossEntropyLoss() 41 | self.hand_dxdymagnitude_loss = nn.MSELoss() 42 | self.hand_handside_loss = nn.BCEWithLogitsLoss() 43 | 44 | # 45 | self._init_weights() 46 | 47 | 48 | 49 | def hand_contactstate_part(self, input, roi_labels, box_info): 50 | contactstate_pred = self.hand_contact_state_layer(input) 51 | contactstate_loss = 0 52 | 53 | if self.training: 54 | for i in range(input.size(0)): 55 | gt_labels = box_info[i, :, 0] # contactstate label 56 | index = roi_labels[i]==2 # if class is hand 57 | if index.sum() > 0: 58 | contactstate_loss_sub = 0.1 * self.hand_contactstate_loss(contactstate_pred[i][index], gt_labels[index].long()) 59 | 60 | if not contactstate_loss: 61 | contactstate_loss = contactstate_loss_sub 62 | else: 63 | contactstate_loss += contactstate_loss_sub 64 | 65 | return contactstate_pred, contactstate_loss 66 | 67 | 68 | 69 | def hand_dxdymagnitude_part(self, input, roi_labels, box_info): 70 | dxdymagnitude_pred = self.hand_dydx_layer(input) 71 | 72 | dxdymagnitude_pred_sub = 0.1 * F.normalize(dxdymagnitude_pred[:,:,1:], p=2, dim=2) 73 | 74 | dxdymagnitude_pred_norm = torch.cat([dxdymagnitude_pred[:,:,0].unsqueeze(-1), dxdymagnitude_pred_sub], dim=2) 75 | dxdymagnitude_loss = 0 76 | 77 | if self.training: 78 | # if 1: 79 | for i in range(input.size(0)): 80 | gt_labels = box_info[i, :, 2:5] # [magnitude, dx, dy] label 81 | index = box_info[i,:,0] > 0 # in-contact 82 | 83 | if index.sum() > 0: 84 | dxdymagnitude_loss_sub = 0.1 * self.hand_dxdymagnitude_loss(dxdymagnitude_pred_norm[i][index], gt_labels[index]) 85 | 86 | if not dxdymagnitude_loss: 87 | dxdymagnitude_loss = dxdymagnitude_loss_sub 88 | else: 89 | dxdymagnitude_loss += dxdymagnitude_loss_sub 90 | 91 | 92 | 93 | return dxdymagnitude_pred_norm, dxdymagnitude_loss 94 | 95 | 96 | 97 | def hand_handside_part(self, input, roi_labels, box_info): 98 | handside_pred = self.hand_lr_layer(input) 99 | handside_loss = 0 100 | 101 | if self.training: 102 | for i in range(input.size(0)): 103 | gt_labels = box_info[i, :, 1] # hand side label 104 | index = roi_labels[i]==2 # if class is hand 105 | if index.sum() > 0: 106 | handside_loss_sub = 0.1 * self.hand_handside_loss(handside_pred[i][index], gt_labels[index].unsqueeze(-1)) 107 | 108 | if not handside_loss: 109 | handside_loss = handside_loss_sub 110 | else: 111 | handside_loss += handside_loss_sub 112 | 113 | return handside_pred, handside_loss 114 | 115 | 116 | 117 | def _init_weights(self): 118 | def normal_init(m, mean, stddev, truncated=False): 119 | """ 120 | weight initalizer: truncated normal and random normal. 121 | """ 122 | # x is a parameter 123 | if truncated: 124 | m.weight.data.normal_().fmod_(2).mul_(stddev).add_(mean) 125 | else: 126 | m.weight.data.normal_(mean, stddev) 127 | m.bias.data.zero_() 128 | 129 | normal_init(self.hand_contact_state_layer[0], 0, 0.01) 130 | normal_init(self.hand_contact_state_layer[3], 0, 0.01) 131 | normal_init(self.hand_dydx_layer, 0, 0.01) 132 | normal_init(self.hand_lr_layer, 0, 0.01) 133 | -------------------------------------------------------------------------------- /lib/model/utils/matching.py: -------------------------------------------------------------------------------- 1 | import numpy as np 2 | import os, pickle 3 | from PIL import Image, ImageDraw, ImageFont 4 | from viz_hand_obj import * 5 | 6 | 7 | def calculate_center(bb): 8 | return [(bb[0] + bb[2])/2, (bb[1] + bb[3])/2] 9 | 10 | def filter_object(obj_dets, hand_dets): 11 | object_cc_list = [] # object center list 12 | for j in range(obj_dets.shape[0]): 13 | object_cc_list.append(calculate_center(obj_dets[j,:4])) 14 | object_cc_list = np.array(object_cc_list) 15 | 16 | img_obj_id = [] # matching list 17 | for i in range(hand_dets.shape[0]): 18 | if hand_dets[i, 5] <= 0: # if hand is non-contact 19 | img_obj_id.append(-1) 20 | continue 21 | else: # hand is in-contact 22 | hand_cc = np.array(calculate_center(hand_dets[i,:4])) # hand center points 23 | point_cc = np.array([(hand_cc[0]+hand_dets[i,6]*10000*hand_dets[i,7]), (hand_cc[1]+hand_dets[i,6]*10000*hand_dets[i,8])]) # extended points (hand center + offset) 24 | dist = np.sum((object_cc_list - point_cc)**2,axis=1) 25 | dist_min = np.argmin(dist) # find the nearest 26 | img_obj_id.append(dist_min) 27 | 28 | return img_obj_id 29 | 30 | 31 | if __name__ == '__main__': 32 | 33 | ############################################################## 34 | # save the detection results in a pickle file, in dictionary format like 35 | 36 | # {"input_image_path_1":{ 37 | # "hand_dets": hand_dets 38 | # ""'obj_dets": obj_dets 39 | # } 40 | # "input_image_path_2":{ 41 | # "hand_dets": hand_dets 42 | # "obj_dets": obj_dets 43 | # } 44 | # .... 45 | # } 46 | 47 | # with img_obj_id generated by filter_object() function, hand_dets and obj_dets can be matched 48 | 49 | ############################################################## 50 | 51 | pickle_path = 'path/to/saved/pickle/file' 52 | save_dir = 'path/to/save' 53 | os.makedirs(save_dir, exist_ok=True) 54 | thresh_hand = 0.5 55 | thresh_obj = 0.5 56 | viz = True 57 | 58 | with open(pickle_path, 'rb') as f: 59 | pickle_info = pickle.load(f) 60 | 61 | for image_path, image_info in pickle_info.items(): 62 | hand_dets = image_info['hand_dets'] 63 | obj_dets = image_info['obj_dets'] 64 | image_name = os.path.split(image_path)[-1] 65 | 66 | if viz: 67 | image = Image.open(image_path).convert("RGBA") 68 | draw = ImageDraw.Draw(image) 69 | font = ImageFont.truetype('times_b.ttf', size=30) 70 | width, height = image.size 71 | 72 | 73 | if (obj_dets is not None) and (hand_dets is not None): 74 | # get matching list 75 | img_obj_id = filter_object(obj_dets, hand_dets) 76 | 77 | # obj 78 | for obj_idx, i in enumerate(range(np.minimum(10, obj_dets.shape[0]))): 79 | obj_bbox = list(int(np.round(x)) for x in obj_dets[i, :4]) 80 | obj_score = obj_dets[i, 4] 81 | 82 | # viz 83 | if viz and obj_score > thresh_obj and i in img_obj_id: # draw obj if > threshold and matched with a hand 84 | image = draw_obj_mask(image, draw, obj_idx, obj_bbox, obj_score, width, height, font) 85 | 86 | 87 | # hand 88 | for hand_idx, i in enumerate(range(np.minimum(10, hand_dets.shape[0]))): 89 | hand_bbox = list(int(np.round(x)) for x in hand_dets[i, :4]) 90 | hand_score = hand_dets[i, 4] 91 | hand_state = hand_dets[i, 5] 92 | hand_vec = hand_dets[i, 6:9] 93 | hand_lr = hand_dets[i, -1] 94 | matched_obj = obj_dets[img_obj_id[i],:4] 95 | 96 | # viz 97 | 98 | if viz and hand_score > thresh_hand: 99 | image = draw_hand_mask(image, draw, hand_idx, hand_bbox, hand_score, hand_lr, hand_state, width, height, font) 100 | if hand_state > 0: # in contact hand 101 | 102 | obj_cc, hand_cc = calculate_center(matched_obj), calculate_center(hand_bbox) 103 | side_idx = int(hand_lr) 104 | draw_line_point(draw, side_idx, (int(hand_cc[0]), int(hand_cc[1])), (int(obj_cc[0]), int(obj_cc[1]))) 105 | 106 | elif hand_dets is not None: 107 | 108 | for hand_idx, i in enumerate(range(np.minimum(10, hand_dets.shape[0]))): 109 | hand_bbox = list(int(np.round(x)) for x in hand_dets[i, :4]) 110 | hand_score = hand_dets[i, 4] 111 | hand_state = hand_dets[i, 5] 112 | hand_lr = hand_dets[i, -1] 113 | 114 | if viz and hand_score > thresh_hand: 115 | image = draw_hand_mask(image, draw, hand_idx, hand_bbox, hand_score, hand_lr, hand_state, width, height, font) 116 | 117 | if viz: 118 | save_path = os.path.join(save_dir, image_name)[:-4] +'_det.png' 119 | image.save(save_path) 120 | 121 | 122 | 123 | -------------------------------------------------------------------------------- /lib/model/csrc/cuda/nms.cu: -------------------------------------------------------------------------------- 1 | // Copyright (c) Facebook, Inc. and its affiliates. All Rights Reserved. 2 | #include 3 | #include 4 | #include 5 | #include 6 | // #include 7 | // #include 8 | 9 | #include 10 | #include 11 | 12 | int const threadsPerBlock = sizeof(unsigned long long) * 8; 13 | 14 | __device__ inline float devIoU(float const * const a, float const * const b) { 15 | float left = max(a[0], b[0]), right = min(a[2], b[2]); 16 | float top = max(a[1], b[1]), bottom = min(a[3], b[3]); 17 | float width = max(right - left + 1, 0.f), height = max(bottom - top + 1, 0.f); 18 | float interS = width * height; 19 | float Sa = (a[2] - a[0] + 1) * (a[3] - a[1] + 1); 20 | float Sb = (b[2] - b[0] + 1) * (b[3] - b[1] + 1); 21 | return interS / (Sa + Sb - interS); 22 | } 23 | 24 | __global__ void nms_kernel(const int n_boxes, const float nms_overlap_thresh, 25 | const float *dev_boxes, unsigned long long *dev_mask) { 26 | const int row_start = blockIdx.y; 27 | const int col_start = blockIdx.x; 28 | 29 | // if (row_start > col_start) return; 30 | 31 | const int row_size = 32 | min(n_boxes - row_start * threadsPerBlock, threadsPerBlock); 33 | const int col_size = 34 | min(n_boxes - col_start * threadsPerBlock, threadsPerBlock); 35 | 36 | __shared__ float block_boxes[threadsPerBlock * 5]; 37 | if (threadIdx.x < col_size) { 38 | block_boxes[threadIdx.x * 5 + 0] = 39 | dev_boxes[(threadsPerBlock * col_start + threadIdx.x) * 5 + 0]; 40 | block_boxes[threadIdx.x * 5 + 1] = 41 | dev_boxes[(threadsPerBlock * col_start + threadIdx.x) * 5 + 1]; 42 | block_boxes[threadIdx.x * 5 + 2] = 43 | dev_boxes[(threadsPerBlock * col_start + threadIdx.x) * 5 + 2]; 44 | block_boxes[threadIdx.x * 5 + 3] = 45 | dev_boxes[(threadsPerBlock * col_start + threadIdx.x) * 5 + 3]; 46 | block_boxes[threadIdx.x * 5 + 4] = 47 | dev_boxes[(threadsPerBlock * col_start + threadIdx.x) * 5 + 4]; 48 | } 49 | __syncthreads(); 50 | 51 | if (threadIdx.x < row_size) { 52 | const int cur_box_idx = threadsPerBlock * row_start + threadIdx.x; 53 | const float *cur_box = dev_boxes + cur_box_idx * 5; 54 | int i = 0; 55 | unsigned long long t = 0; 56 | int start = 0; 57 | if (row_start == col_start) { 58 | start = threadIdx.x + 1; 59 | } 60 | for (i = start; i < col_size; i++) { 61 | if (devIoU(cur_box, block_boxes + i * 5) > nms_overlap_thresh) { 62 | t |= 1ULL << i; 63 | } 64 | } 65 | const int col_blocks = at::ceil_div(n_boxes, threadsPerBlock); 66 | dev_mask[cur_box_idx * col_blocks + col_start] = t; 67 | } 68 | } 69 | 70 | // boxes is a N x 5 tensor 71 | at::Tensor nms_cuda(const at::Tensor boxes, float nms_overlap_thresh) { 72 | using scalar_t = float; 73 | AT_ASSERTM(boxes.type().is_cuda(), "boxes must be a CUDA tensor"); 74 | auto scores = boxes.select(1, 4); 75 | auto order_t = std::get<1>(scores.sort(0, /* descending=*/true)); 76 | auto boxes_sorted = boxes.index_select(0, order_t); 77 | 78 | int boxes_num = boxes.size(0); 79 | 80 | const int col_blocks = at::ceil_div(boxes_num, threadsPerBlock); 81 | 82 | scalar_t* boxes_dev = boxes_sorted.data(); 83 | 84 | // THCState *state = at::globalContext().lazyInitCUDA(); // TODO replace with getTHCState 85 | 86 | unsigned long long* mask_dev = NULL; 87 | //THCudaCheck(THCudaMalloc(state, (void**) &mask_dev, 88 | // boxes_num * col_blocks * sizeof(unsigned long long))); 89 | 90 | // mask_dev = (unsigned long long*) THCudaMalloc(state, boxes_num * col_blocks * sizeof(unsigned long long)); 91 | mask_dev = (unsigned long long*) c10::cuda::CUDACachingAllocator::raw_alloc(boxes_num * col_blocks * sizeof(unsigned long long)); 92 | dim3 blocks(at::ceil_div(boxes_num, threadsPerBlock), 93 | at::ceil_div(boxes_num, threadsPerBlock)); 94 | dim3 threads(threadsPerBlock); 95 | nms_kernel<<>>(boxes_num, 96 | nms_overlap_thresh, 97 | boxes_dev, 98 | mask_dev); 99 | 100 | std::vector mask_host(boxes_num * col_blocks); 101 | C10_CUDA_CHECK(cudaMemcpy(&mask_host[0], 102 | mask_dev, 103 | sizeof(unsigned long long) * boxes_num * col_blocks, 104 | cudaMemcpyDeviceToHost)); 105 | 106 | std::vector remv(col_blocks); 107 | memset(&remv[0], 0, sizeof(unsigned long long) * col_blocks); 108 | 109 | at::Tensor keep = at::empty({boxes_num}, boxes.options().dtype(at::kLong).device(at::kCPU)); 110 | int64_t* keep_out = keep.data(); 111 | 112 | int num_to_keep = 0; 113 | for (int i = 0; i < boxes_num; i++) { 114 | int nblock = i / threadsPerBlock; 115 | int inblock = i % threadsPerBlock; 116 | 117 | if (!(remv[nblock] & (1ULL << inblock))) { 118 | keep_out[num_to_keep++] = i; 119 | unsigned long long *p = &mask_host[0] + i * col_blocks; 120 | for (int j = nblock; j < col_blocks; j++) { 121 | remv[j] |= p[j]; 122 | } 123 | } 124 | } 125 | 126 | c10::cuda::CUDACachingAllocator::raw_delete(mask_dev); 127 | // THCudaFree(state, mask_dev); 128 | // TODO improve this part 129 | return std::get<0>(order_t.index({ 130 | keep.narrow(/*dim=*/0, /*start=*/0, /*length=*/num_to_keep).to( 131 | order_t.device(), keep.scalar_type()) 132 | }).sort(0, false)); 133 | } -------------------------------------------------------------------------------- /lib/model/nms/nms_kernel.cu: -------------------------------------------------------------------------------- 1 | // ------------------------------------------------------------------ 2 | // Faster R-CNN 3 | // Copyright (c) 2015 Microsoft 4 | // Licensed under The MIT License [see fast-rcnn/LICENSE for details] 5 | // Written by Shaoqing Ren 6 | // ------------------------------------------------------------------ 7 | 8 | #include "gpu_nms.hpp" 9 | #include 10 | #include 11 | 12 | #define CUDA_CHECK(condition) \ 13 | /* Code block avoids redefinition of cudaError_t error */ \ 14 | do { \ 15 | cudaError_t error = condition; \ 16 | if (error != cudaSuccess) { \ 17 | std::cout << cudaGetErrorString(error) << std::endl; \ 18 | } \ 19 | } while (0) 20 | 21 | #define DIVUP(m,n) ((m) / (n) + ((m) % (n) > 0)) 22 | int const threadsPerBlock = sizeof(unsigned long long) * 8; 23 | 24 | __device__ inline float devIoU(float const * const a, float const * const b) { 25 | float left = max(a[0], b[0]), right = min(a[2], b[2]); 26 | float top = max(a[1], b[1]), bottom = min(a[3], b[3]); 27 | float width = max(right - left + 1, 0.f), height = max(bottom - top + 1, 0.f); 28 | float interS = width * height; 29 | float Sa = (a[2] - a[0] + 1) * (a[3] - a[1] + 1); 30 | float Sb = (b[2] - b[0] + 1) * (b[3] - b[1] + 1); 31 | return interS / (Sa + Sb - interS); 32 | } 33 | 34 | __global__ void nms_kernel(const int n_boxes, const float nms_overlap_thresh, 35 | const float *dev_boxes, unsigned long long *dev_mask) { 36 | const int row_start = blockIdx.y; 37 | const int col_start = blockIdx.x; 38 | 39 | // if (row_start > col_start) return; 40 | 41 | const int row_size = 42 | min(n_boxes - row_start * threadsPerBlock, threadsPerBlock); 43 | const int col_size = 44 | min(n_boxes - col_start * threadsPerBlock, threadsPerBlock); 45 | 46 | __shared__ float block_boxes[threadsPerBlock * 5]; 47 | if (threadIdx.x < col_size) { 48 | block_boxes[threadIdx.x * 5 + 0] = 49 | dev_boxes[(threadsPerBlock * col_start + threadIdx.x) * 5 + 0]; 50 | block_boxes[threadIdx.x * 5 + 1] = 51 | dev_boxes[(threadsPerBlock * col_start + threadIdx.x) * 5 + 1]; 52 | block_boxes[threadIdx.x * 5 + 2] = 53 | dev_boxes[(threadsPerBlock * col_start + threadIdx.x) * 5 + 2]; 54 | block_boxes[threadIdx.x * 5 + 3] = 55 | dev_boxes[(threadsPerBlock * col_start + threadIdx.x) * 5 + 3]; 56 | block_boxes[threadIdx.x * 5 + 4] = 57 | dev_boxes[(threadsPerBlock * col_start + threadIdx.x) * 5 + 4]; 58 | } 59 | __syncthreads(); 60 | 61 | if (threadIdx.x < row_size) { 62 | const int cur_box_idx = threadsPerBlock * row_start + threadIdx.x; 63 | const float *cur_box = dev_boxes + cur_box_idx * 5; 64 | int i = 0; 65 | unsigned long long t = 0; 66 | int start = 0; 67 | if (row_start == col_start) { 68 | start = threadIdx.x + 1; 69 | } 70 | for (i = start; i < col_size; i++) { 71 | if (devIoU(cur_box, block_boxes + i * 5) > nms_overlap_thresh) { 72 | t |= 1ULL << i; 73 | } 74 | } 75 | const int col_blocks = DIVUP(n_boxes, threadsPerBlock); 76 | dev_mask[cur_box_idx * col_blocks + col_start] = t; 77 | } 78 | } 79 | 80 | void _set_device(int device_id) { 81 | int current_device; 82 | CUDA_CHECK(cudaGetDevice(¤t_device)); 83 | if (current_device == device_id) { 84 | return; 85 | } 86 | // The call to cudaSetDevice must come before any calls to Get, which 87 | // may perform initialization using the GPU. 88 | CUDA_CHECK(cudaSetDevice(device_id)); 89 | } 90 | 91 | void _nms(int* keep_out, int* num_out, const float* boxes_host, int boxes_num, 92 | int boxes_dim, float nms_overlap_thresh, int device_id) { 93 | _set_device(device_id); 94 | 95 | float* boxes_dev = NULL; 96 | unsigned long long* mask_dev = NULL; 97 | 98 | const int col_blocks = DIVUP(boxes_num, threadsPerBlock); 99 | 100 | CUDA_CHECK(cudaMalloc(&boxes_dev, 101 | boxes_num * boxes_dim * sizeof(float))); 102 | CUDA_CHECK(cudaMemcpy(boxes_dev, 103 | boxes_host, 104 | boxes_num * boxes_dim * sizeof(float), 105 | cudaMemcpyHostToDevice)); 106 | 107 | CUDA_CHECK(cudaMalloc(&mask_dev, 108 | boxes_num * col_blocks * sizeof(unsigned long long))); 109 | 110 | dim3 blocks(DIVUP(boxes_num, threadsPerBlock), 111 | DIVUP(boxes_num, threadsPerBlock)); 112 | dim3 threads(threadsPerBlock); 113 | nms_kernel<<>>(boxes_num, 114 | nms_overlap_thresh, 115 | boxes_dev, 116 | mask_dev); 117 | 118 | std::vector mask_host(boxes_num * col_blocks); 119 | CUDA_CHECK(cudaMemcpy(&mask_host[0], 120 | mask_dev, 121 | sizeof(unsigned long long) * boxes_num * col_blocks, 122 | cudaMemcpyDeviceToHost)); 123 | 124 | std::vector remv(col_blocks); 125 | memset(&remv[0], 0, sizeof(unsigned long long) * col_blocks); 126 | 127 | int num_to_keep = 0; 128 | for (int i = 0; i < boxes_num; i++) { 129 | int nblock = i / threadsPerBlock; 130 | int inblock = i % threadsPerBlock; 131 | 132 | if (!(remv[nblock] & (1ULL << inblock))) { 133 | keep_out[num_to_keep++] = i; 134 | unsigned long long *p = &mask_host[0] + i * col_blocks; 135 | for (int j = nblock; j < col_blocks; j++) { 136 | remv[j] |= p[j]; 137 | } 138 | } 139 | } 140 | *num_out = num_to_keep; 141 | 142 | CUDA_CHECK(cudaFree(boxes_dev)); 143 | CUDA_CHECK(cudaFree(mask_dev)); 144 | } 145 | -------------------------------------------------------------------------------- /lib/model/nms/src/nms_cuda_kernel.cu: -------------------------------------------------------------------------------- 1 | // ------------------------------------------------------------------ 2 | // Faster R-CNN 3 | // Copyright (c) 2015 Microsoft 4 | // Licensed under The MIT License [see fast-rcnn/LICENSE for details] 5 | // Written by Shaoqing Ren 6 | // ------------------------------------------------------------------ 7 | 8 | #include 9 | #include 10 | #include 11 | #include 12 | #include "nms_cuda_kernel.h" 13 | 14 | #define CUDA_WARN(XXX) \ 15 | do { if (XXX != cudaSuccess) std::cout << "CUDA Error: " << \ 16 | cudaGetErrorString(XXX) << ", at line " << __LINE__ \ 17 | << std::endl; cudaDeviceSynchronize(); } while (0) 18 | 19 | #define CUDA_CHECK(condition) \ 20 | /* Code block avoids redefinition of cudaError_t error */ \ 21 | do { \ 22 | cudaError_t error = condition; \ 23 | if (error != cudaSuccess) { \ 24 | std::cout << cudaGetErrorString(error) << std::endl; \ 25 | } \ 26 | } while (0) 27 | 28 | #define DIVUP(m,n) ((m) / (n) + ((m) % (n) > 0)) 29 | int const threadsPerBlock = sizeof(unsigned long long) * 8; 30 | 31 | __device__ inline float devIoU(float const * const a, float const * const b) { 32 | float left = max(a[0], b[0]), right = min(a[2], b[2]); 33 | float top = max(a[1], b[1]), bottom = min(a[3], b[3]); 34 | float width = max(right - left + 1, 0.f), height = max(bottom - top + 1, 0.f); 35 | float interS = width * height; 36 | float Sa = (a[2] - a[0] + 1) * (a[3] - a[1] + 1); 37 | float Sb = (b[2] - b[0] + 1) * (b[3] - b[1] + 1); 38 | return interS / (Sa + Sb - interS); 39 | } 40 | 41 | __global__ void nms_kernel(int n_boxes, float nms_overlap_thresh, 42 | float *dev_boxes, unsigned long long *dev_mask) { 43 | const int row_start = blockIdx.y; 44 | const int col_start = blockIdx.x; 45 | 46 | // if (row_start > col_start) return; 47 | 48 | const int row_size = 49 | min(n_boxes - row_start * threadsPerBlock, threadsPerBlock); 50 | const int col_size = 51 | min(n_boxes - col_start * threadsPerBlock, threadsPerBlock); 52 | 53 | __shared__ float block_boxes[threadsPerBlock * 5]; 54 | if (threadIdx.x < col_size) { 55 | block_boxes[threadIdx.x * 5 + 0] = 56 | dev_boxes[(threadsPerBlock * col_start + threadIdx.x) * 5 + 0]; 57 | block_boxes[threadIdx.x * 5 + 1] = 58 | dev_boxes[(threadsPerBlock * col_start + threadIdx.x) * 5 + 1]; 59 | block_boxes[threadIdx.x * 5 + 2] = 60 | dev_boxes[(threadsPerBlock * col_start + threadIdx.x) * 5 + 2]; 61 | block_boxes[threadIdx.x * 5 + 3] = 62 | dev_boxes[(threadsPerBlock * col_start + threadIdx.x) * 5 + 3]; 63 | block_boxes[threadIdx.x * 5 + 4] = 64 | dev_boxes[(threadsPerBlock * col_start + threadIdx.x) * 5 + 4]; 65 | } 66 | __syncthreads(); 67 | 68 | if (threadIdx.x < row_size) { 69 | const int cur_box_idx = threadsPerBlock * row_start + threadIdx.x; 70 | const float *cur_box = dev_boxes + cur_box_idx * 5; 71 | int i = 0; 72 | unsigned long long t = 0; 73 | int start = 0; 74 | if (row_start == col_start) { 75 | start = threadIdx.x + 1; 76 | } 77 | for (i = start; i < col_size; i++) { 78 | if (devIoU(cur_box, block_boxes + i * 5) > nms_overlap_thresh) { 79 | t |= 1ULL << i; 80 | } 81 | } 82 | const int col_blocks = DIVUP(n_boxes, threadsPerBlock); 83 | dev_mask[cur_box_idx * col_blocks + col_start] = t; 84 | } 85 | } 86 | 87 | void nms_cuda_compute(int* keep_out, int *num_out, float* boxes_host, int boxes_num, 88 | int boxes_dim, float nms_overlap_thresh) { 89 | 90 | float* boxes_dev = NULL; 91 | unsigned long long* mask_dev = NULL; 92 | 93 | const int col_blocks = DIVUP(boxes_num, threadsPerBlock); 94 | 95 | CUDA_CHECK(cudaMalloc(&boxes_dev, 96 | boxes_num * boxes_dim * sizeof(float))); 97 | CUDA_CHECK(cudaMemcpy(boxes_dev, 98 | boxes_host, 99 | boxes_num * boxes_dim * sizeof(float), 100 | cudaMemcpyHostToDevice)); 101 | 102 | CUDA_CHECK(cudaMalloc(&mask_dev, 103 | boxes_num * col_blocks * sizeof(unsigned long long))); 104 | 105 | dim3 blocks(DIVUP(boxes_num, threadsPerBlock), 106 | DIVUP(boxes_num, threadsPerBlock)); 107 | dim3 threads(threadsPerBlock); 108 | 109 | // printf("i am at line %d\n", boxes_num); 110 | // printf("i am at line %d\n", boxes_dim); 111 | 112 | nms_kernel<<>>(boxes_num, 113 | nms_overlap_thresh, 114 | boxes_dev, 115 | mask_dev); 116 | 117 | std::vector mask_host(boxes_num * col_blocks); 118 | CUDA_CHECK(cudaMemcpy(&mask_host[0], 119 | mask_dev, 120 | sizeof(unsigned long long) * boxes_num * col_blocks, 121 | cudaMemcpyDeviceToHost)); 122 | 123 | std::vector remv(col_blocks); 124 | memset(&remv[0], 0, sizeof(unsigned long long) * col_blocks); 125 | 126 | // we need to create a memory for keep_out on cpu 127 | // otherwise, the following code cannot run 128 | 129 | int* keep_out_cpu = new int[boxes_num]; 130 | 131 | int num_to_keep = 0; 132 | for (int i = 0; i < boxes_num; i++) { 133 | int nblock = i / threadsPerBlock; 134 | int inblock = i % threadsPerBlock; 135 | 136 | if (!(remv[nblock] & (1ULL << inblock))) { 137 | // orignal: keep_out[num_to_keep++] = i; 138 | keep_out_cpu[num_to_keep++] = i; 139 | unsigned long long *p = &mask_host[0] + i * col_blocks; 140 | for (int j = nblock; j < col_blocks; j++) { 141 | remv[j] |= p[j]; 142 | } 143 | } 144 | } 145 | 146 | // copy keep_out_cpu to keep_out on gpu 147 | CUDA_WARN(cudaMemcpy(keep_out, keep_out_cpu, boxes_num * sizeof(int),cudaMemcpyHostToDevice)); 148 | 149 | // *num_out = num_to_keep; 150 | 151 | // original: *num_out = num_to_keep; 152 | // copy num_to_keep to num_out on gpu 153 | 154 | CUDA_WARN(cudaMemcpy(num_out, &num_to_keep, 1 * sizeof(int),cudaMemcpyHostToDevice)); 155 | 156 | // release cuda memory 157 | CUDA_CHECK(cudaFree(boxes_dev)); 158 | CUDA_CHECK(cudaFree(mask_dev)); 159 | // release cpu memory 160 | delete []keep_out_cpu; 161 | } 162 | -------------------------------------------------------------------------------- /lib/datasets/voc_eval_raw.py: -------------------------------------------------------------------------------- 1 | # -------------------------------------------------------- 2 | # Fast/er R-CNN 3 | # Licensed under The MIT License [see LICENSE for details] 4 | # Written by Bharath Hariharan 5 | # -------------------------------------------------------- 6 | from __future__ import absolute_import 7 | from __future__ import division 8 | from __future__ import print_function 9 | 10 | import xml.etree.ElementTree as ET 11 | import os 12 | import pickle 13 | import numpy as np 14 | 15 | def parse_rec(filename): 16 | """ Parse a PASCAL VOC xml file """ 17 | tree = ET.parse(filename) 18 | objects = [] 19 | for obj in tree.findall('object'): 20 | obj_struct = {} 21 | obj_struct['name'] = obj.find('name').text 22 | obj_struct['pose'] = obj.find('pose').text 23 | obj_struct['truncated'] = int(obj.find('truncated').text) 24 | obj_struct['difficult'] = int(obj.find('difficult').text) 25 | bbox = obj.find('bndbox') 26 | obj_struct['bbox'] = [int(bbox.find('xmin').text), 27 | int(bbox.find('ymin').text), 28 | int(bbox.find('xmax').text), 29 | int(bbox.find('ymax').text)] 30 | objects.append(obj_struct) 31 | 32 | return objects 33 | 34 | 35 | def voc_ap(rec, prec, use_07_metric=False): 36 | """ ap = voc_ap(rec, prec, [use_07_metric]) 37 | Compute VOC AP given precision and recall. 38 | If use_07_metric is true, uses the 39 | VOC 07 11 point method (default:False). 40 | """ 41 | if use_07_metric: 42 | # 11 point metric 43 | ap = 0. 44 | for t in np.arange(0., 1.1, 0.1): 45 | if np.sum(rec >= t) == 0: 46 | p = 0 47 | else: 48 | p = np.max(prec[rec >= t]) 49 | ap = ap + p / 11. 50 | else: 51 | # correct AP calculation 52 | # first append sentinel values at the end 53 | mrec = np.concatenate(([0.], rec, [1.])) 54 | mpre = np.concatenate(([0.], prec, [0.])) 55 | 56 | # compute the precision envelope 57 | for i in range(mpre.size - 1, 0, -1): 58 | mpre[i - 1] = np.maximum(mpre[i - 1], mpre[i]) 59 | 60 | # to calculate area under PR curve, look for points 61 | # where X axis (recall) changes value 62 | i = np.where(mrec[1:] != mrec[:-1])[0] 63 | 64 | # and sum (\Delta recall) * prec 65 | ap = np.sum((mrec[i + 1] - mrec[i]) * mpre[i + 1]) 66 | return ap 67 | 68 | 69 | def voc_eval(detpath, 70 | annopath, 71 | imagesetfile, 72 | classname, 73 | cachedir, 74 | ovthresh=0.7, 75 | use_07_metric=False): 76 | """rec, prec, ap = voc_eval(detpath, 77 | annopath, 78 | imagesetfile, 79 | classname, 80 | [ovthresh], 81 | [use_07_metric]) 82 | 83 | Top level function that does the PASCAL VOC evaluation. 84 | 85 | detpath: Path to detections 86 | detpath.format(classname) should produce the detection results file. 87 | annopath: Path to annotations 88 | annopath.format(imagename) should be the xml annotations file. 89 | imagesetfile: Text file containing the list of images, one image per line. 90 | classname: Category name (duh) 91 | cachedir: Directory for caching the annotations 92 | [ovthresh]: Overlap threshold (default = 0.5) 93 | [use_07_metric]: Whether to use VOC07's 11 point AP computation 94 | (default False) 95 | """ 96 | # assumes detections are in detpath.format(classname) 97 | # assumes annotations are in annopath.format(imagename) 98 | # assumes imagesetfile is a text file with each line an image name 99 | # cachedir caches the annotations in a pickle file 100 | 101 | print(f'\n\n thd = {ovthresh}\n\n') 102 | 103 | # first load gt 104 | if not os.path.isdir(cachedir): 105 | os.mkdir(cachedir) 106 | cachefile = os.path.join(cachedir, '%s_annots.pkl' % imagesetfile) 107 | # read list of images 108 | with open(imagesetfile, 'r') as f: 109 | lines = f.readlines() 110 | imagenames = [x.strip() for x in lines] 111 | 112 | if not os.path.isfile(cachefile): 113 | # load annotations 114 | recs = {} 115 | for i, imagename in enumerate(imagenames): 116 | recs[imagename] = parse_rec(annopath.format(imagename)) 117 | if i % 100 == 0: 118 | print('Reading annotation for {:d}/{:d}'.format( 119 | i + 1, len(imagenames))) 120 | # save 121 | print('Saving cached annotations to {:s}'.format(cachefile)) 122 | with open(cachefile, 'wb') as f: 123 | pickle.dump(recs, f) 124 | else: 125 | # load 126 | with open(cachefile, 'rb') as f: 127 | try: 128 | recs = pickle.load(f) 129 | except: 130 | recs = pickle.load(f, encoding='bytes') 131 | 132 | # extract gt objects for this class 133 | class_recs = {} 134 | npos = 0 135 | for imagename in imagenames: 136 | R = [obj for obj in recs[imagename] if obj['name'] == classname] 137 | bbox = np.array([x['bbox'] for x in R]) 138 | difficult = np.array([x['difficult'] for x in R]).astype(np.bool) 139 | det = [False] * len(R) 140 | npos = npos + sum(~difficult) 141 | class_recs[imagename] = {'bbox': bbox, 142 | 'difficult': difficult, 143 | 'det': det} 144 | 145 | # read dets 146 | detfile = detpath.format(classname) 147 | with open(detfile, 'r') as f: 148 | lines = f.readlines() 149 | 150 | splitlines = [x.strip().split(' ') for x in lines] 151 | image_ids = [x[0] for x in splitlines] 152 | confidence = np.array([float(x[1]) for x in splitlines]) 153 | BB = np.array([[float(z) for z in x[2:]] for x in splitlines]) 154 | 155 | nd = len(image_ids) 156 | tp = np.zeros(nd) 157 | fp = np.zeros(nd) 158 | 159 | if BB.shape[0] > 0: 160 | # sort by confidence 161 | sorted_ind = np.argsort(-confidence) 162 | sorted_scores = np.sort(-confidence) 163 | BB = BB[sorted_ind, :] 164 | image_ids = [image_ids[x] for x in sorted_ind] 165 | 166 | # go down dets and mark TPs and FPs 167 | for d in range(nd): 168 | R = class_recs[image_ids[d]] 169 | bb = BB[d, :].astype(float) 170 | ovmax = -np.inf 171 | BBGT = R['bbox'].astype(float) 172 | 173 | if BBGT.size > 0: 174 | # compute overlaps 175 | # intersection 176 | ixmin = np.maximum(BBGT[:, 0], bb[0]) 177 | iymin = np.maximum(BBGT[:, 1], bb[1]) 178 | ixmax = np.minimum(BBGT[:, 2], bb[2]) 179 | iymax = np.minimum(BBGT[:, 3], bb[3]) 180 | iw = np.maximum(ixmax - ixmin + 1., 0.) 181 | ih = np.maximum(iymax - iymin + 1., 0.) 182 | inters = iw * ih 183 | 184 | # union 185 | uni = ((bb[2] - bb[0] + 1.) * (bb[3] - bb[1] + 1.) + 186 | (BBGT[:, 2] - BBGT[:, 0] + 1.) * 187 | (BBGT[:, 3] - BBGT[:, 1] + 1.) - inters) 188 | 189 | overlaps = inters / uni 190 | ovmax = np.max(overlaps) 191 | jmax = np.argmax(overlaps) 192 | 193 | if ovmax > ovthresh: 194 | if not R['difficult'][jmax]: 195 | if not R['det'][jmax]: 196 | tp[d] = 1. 197 | R['det'][jmax] = 1 198 | else: 199 | fp[d] = 1. 200 | else: 201 | fp[d] = 1. 202 | 203 | # compute precision recall 204 | fp = np.cumsum(fp) 205 | tp = np.cumsum(tp) 206 | rec = tp / float(npos) 207 | # avoid divide by zero in case the first detection matches a difficult 208 | # ground truth 209 | prec = tp / np.maximum(tp + fp, np.finfo(np.float64).eps) 210 | ap = voc_ap(rec, prec, use_07_metric) 211 | 212 | 213 | return rec, prec, ap 214 | -------------------------------------------------------------------------------- /lib/model/rpn/proposal_layer.py: -------------------------------------------------------------------------------- 1 | from __future__ import absolute_import 2 | # -------------------------------------------------------- 3 | # Faster R-CNN 4 | # Copyright (c) 2015 Microsoft 5 | # Licensed under The MIT License [see LICENSE for details] 6 | # Written by Ross Girshick and Sean Bell 7 | # -------------------------------------------------------- 8 | # -------------------------------------------------------- 9 | # Reorganized and modified by Jianwei Yang and Jiasen Lu 10 | # -------------------------------------------------------- 11 | 12 | import torch 13 | import torch.nn as nn 14 | import numpy as np 15 | import math 16 | import yaml 17 | from model.utils.config import cfg 18 | from .generate_anchors import generate_anchors 19 | from .bbox_transform import bbox_transform_inv, clip_boxes, clip_boxes_batch 20 | # from model.nms.nms_wrapper import nms 21 | from model.roi_layers import nms 22 | import pdb 23 | 24 | DEBUG = False 25 | 26 | class _ProposalLayer(nn.Module): 27 | """ 28 | Outputs object detection proposals by applying estimated bounding-box 29 | transformations to a set of regular boxes (called "anchors"). 30 | """ 31 | 32 | def __init__(self, feat_stride, scales, ratios): 33 | super(_ProposalLayer, self).__init__() 34 | 35 | self._feat_stride = feat_stride 36 | self._anchors = torch.from_numpy(generate_anchors(scales=np.array(scales), 37 | ratios=np.array(ratios))).float() 38 | self._num_anchors = self._anchors.size(0) 39 | 40 | # rois blob: holds R regions of interest, each is a 5-tuple 41 | # (n, x1, y1, x2, y2) specifying an image batch index n and a 42 | # rectangle (x1, y1, x2, y2) 43 | # top[0].reshape(1, 5) 44 | # 45 | # # scores blob: holds scores for R regions of interest 46 | # if len(top) > 1: 47 | # top[1].reshape(1, 1, 1, 1) 48 | 49 | def forward(self, input): 50 | 51 | # Algorithm: 52 | # 53 | # for each (H, W) location i 54 | # generate A anchor boxes centered on cell i 55 | # apply predicted bbox deltas at cell i to each of the A anchors 56 | # clip predicted boxes to image 57 | # remove predicted boxes with either height or width < threshold 58 | # sort all (proposal, score) pairs by score from highest to lowest 59 | # take top pre_nms_topN proposals before NMS 60 | # apply NMS with threshold 0.7 to remaining proposals 61 | # take after_nms_topN proposals after NMS 62 | # return the top proposals (-> RoIs top, scores top) 63 | 64 | 65 | # the first set of _num_anchors channels are bg probs 66 | # the second set are the fg probs 67 | scores = input[0][:, self._num_anchors:, :, :] 68 | bbox_deltas = input[1] 69 | im_info = input[2] 70 | cfg_key = input[3] 71 | 72 | pre_nms_topN = cfg[cfg_key].RPN_PRE_NMS_TOP_N 73 | post_nms_topN = cfg[cfg_key].RPN_POST_NMS_TOP_N 74 | nms_thresh = cfg[cfg_key].RPN_NMS_THRESH 75 | min_size = cfg[cfg_key].RPN_MIN_SIZE 76 | 77 | batch_size = bbox_deltas.size(0) 78 | 79 | feat_height, feat_width = scores.size(2), scores.size(3) 80 | shift_x = np.arange(0, feat_width) * self._feat_stride 81 | shift_y = np.arange(0, feat_height) * self._feat_stride 82 | shift_x, shift_y = np.meshgrid(shift_x, shift_y) 83 | shifts = torch.from_numpy(np.vstack((shift_x.ravel(), shift_y.ravel(), 84 | shift_x.ravel(), shift_y.ravel())).transpose()) 85 | shifts = shifts.contiguous().type_as(scores).float() 86 | 87 | A = self._num_anchors 88 | K = shifts.size(0) 89 | 90 | self._anchors = self._anchors.type_as(scores) 91 | # anchors = self._anchors.view(1, A, 4) + shifts.view(1, K, 4).permute(1, 0, 2).contiguous() 92 | anchors = self._anchors.view(1, A, 4) + shifts.view(K, 1, 4) 93 | anchors = anchors.view(1, K * A, 4).expand(batch_size, K * A, 4) 94 | 95 | # Transpose and reshape predicted bbox transformations to get them 96 | # into the same order as the anchors: 97 | 98 | bbox_deltas = bbox_deltas.permute(0, 2, 3, 1).contiguous() 99 | bbox_deltas = bbox_deltas.view(batch_size, -1, 4) 100 | 101 | # Same story for the scores: 102 | scores = scores.permute(0, 2, 3, 1).contiguous() 103 | scores = scores.view(batch_size, -1) 104 | 105 | # Convert anchors into proposals via bbox transformations 106 | proposals = bbox_transform_inv(anchors, bbox_deltas, batch_size) 107 | 108 | # 2. clip predicted boxes to image 109 | proposals = clip_boxes(proposals, im_info, batch_size) 110 | # proposals = clip_boxes_batch(proposals, im_info, batch_size) 111 | 112 | # assign the score to 0 if it's non keep. 113 | # keep = self._filter_boxes(proposals, min_size * im_info[:, 2]) 114 | 115 | # trim keep index to make it euqal over batch 116 | # keep_idx = torch.cat(tuple(keep_idx), 0) 117 | 118 | # scores_keep = scores.view(-1)[keep_idx].view(batch_size, trim_size) 119 | # proposals_keep = proposals.view(-1, 4)[keep_idx, :].contiguous().view(batch_size, trim_size, 4) 120 | 121 | # _, order = torch.sort(scores_keep, 1, True) 122 | 123 | scores_keep = scores 124 | proposals_keep = proposals 125 | _, order = torch.sort(scores_keep, 1, True) 126 | 127 | output = scores.new(batch_size, post_nms_topN, 5).zero_() 128 | for i in range(batch_size): 129 | # # 3. remove predicted boxes with either height or width < threshold 130 | # # (NOTE: convert min_size to input image scale stored in im_info[2]) 131 | proposals_single = proposals_keep[i] 132 | scores_single = scores_keep[i] 133 | 134 | # # 4. sort all (proposal, score) pairs by score from highest to lowest 135 | # # 5. take top pre_nms_topN (e.g. 6000) 136 | order_single = order[i] 137 | 138 | if pre_nms_topN > 0 and pre_nms_topN < scores_keep.numel(): 139 | order_single = order_single[:pre_nms_topN] 140 | 141 | proposals_single = proposals_single[order_single, :] 142 | scores_single = scores_single[order_single].view(-1,1) 143 | 144 | # 6. apply nms (e.g. threshold = 0.7) 145 | # 7. take after_nms_topN (e.g. 300) 146 | # 8. return the top proposals (-> RoIs top) 147 | keep_idx_i = nms(proposals_single, scores_single.squeeze(1), nms_thresh) 148 | keep_idx_i = keep_idx_i.long().view(-1) 149 | 150 | if post_nms_topN > 0: 151 | keep_idx_i = keep_idx_i[:post_nms_topN] 152 | proposals_single = proposals_single[keep_idx_i, :] 153 | scores_single = scores_single[keep_idx_i, :] 154 | 155 | # padding 0 at the end. 156 | num_proposal = proposals_single.size(0) 157 | output[i,:,0] = i 158 | output[i,:num_proposal,1:] = proposals_single 159 | 160 | return output 161 | 162 | def backward(self, top, propagate_down, bottom): 163 | """This layer does not propagate gradients.""" 164 | pass 165 | 166 | def reshape(self, bottom, top): 167 | """Reshaping happens during the call to forward.""" 168 | pass 169 | 170 | def _filter_boxes(self, boxes, min_size): 171 | """Remove all boxes with any side smaller than min_size.""" 172 | ws = boxes[:, :, 2] - boxes[:, :, 0] + 1 173 | hs = boxes[:, :, 3] - boxes[:, :, 1] + 1 174 | keep = ((ws >= min_size.view(-1,1).expand_as(ws)) & (hs >= min_size.view(-1,1).expand_as(hs))) 175 | return keep 176 | -------------------------------------------------------------------------------- /lib/model/faster_rcnn/faster_rcnn.py: -------------------------------------------------------------------------------- 1 | import random 2 | import torch 3 | import torch.nn as nn 4 | import torch.nn.functional as F 5 | from torch.autograd import Variable 6 | import torchvision.models as models 7 | from torch.autograd import Variable 8 | import numpy as np 9 | from model.utils.config import cfg 10 | from model.rpn.rpn import _RPN 11 | from model.extension_layers import extension_layers 12 | 13 | 14 | from model.roi_layers import ROIAlign, ROIPool 15 | 16 | # from model.roi_pooling.modules.roi_pool import _RoIPooling 17 | # from model.roi_align.modules.roi_align import RoIAlignAvg 18 | 19 | from model.rpn.proposal_target_layer_cascade import _ProposalTargetLayer 20 | import time 21 | import pdb 22 | from model.utils.net_utils import _smooth_l1_loss, _crop_pool_layer, _affine_grid_gen, _affine_theta 23 | 24 | class _fasterRCNN(nn.Module): 25 | """ faster RCNN """ 26 | def __init__(self, classes, class_agnostic): 27 | super(_fasterRCNN, self).__init__() 28 | self.classes = classes 29 | self.n_classes = len(classes) 30 | self.class_agnostic = class_agnostic 31 | # loss 32 | self.RCNN_loss_cls = 0 33 | self.RCNN_loss_bbox = 0 34 | 35 | # define rpn 36 | self.RCNN_rpn = _RPN(self.dout_base_model) 37 | self.RCNN_proposal_target = _ProposalTargetLayer(self.n_classes) 38 | 39 | # self.RCNN_roi_pool = _RoIPooling(cfg.POOLING_SIZE, cfg.POOLING_SIZE, 1.0/16.0) 40 | # self.RCNN_roi_align = RoIAlignAvg(cfg.POOLING_SIZE, cfg.POOLING_SIZE, 1.0/16.0) 41 | 42 | self.RCNN_roi_pool = ROIPool((cfg.POOLING_SIZE, cfg.POOLING_SIZE), 1.0/16.0) 43 | self.RCNN_roi_align = ROIAlign((cfg.POOLING_SIZE, cfg.POOLING_SIZE), 1.0/16.0, 0) 44 | self.extension_layer = extension_layers.extension_layer() 45 | 46 | def forward(self, im_data, im_info, gt_boxes, num_boxes, box_info): 47 | batch_size = im_data.size(0) 48 | 49 | im_info = im_info.data 50 | gt_boxes = gt_boxes.data 51 | num_boxes = num_boxes.data 52 | box_info = box_info.data 53 | 54 | # feed image data to base model to obtain base feature map 55 | base_feat = self.RCNN_base(im_data) 56 | # feed base feature map tp RPN to obtain rois 57 | rois, rpn_loss_cls, rpn_loss_bbox = self.RCNN_rpn(base_feat, im_info, gt_boxes, num_boxes) 58 | 59 | 60 | # if it is training phrase, then use ground trubut bboxes for refining 61 | if self.training: 62 | roi_data = self.RCNN_proposal_target(rois, gt_boxes, num_boxes, box_info) 63 | rois, rois_label, rois_target, rois_inside_ws, rois_outside_ws, box_info = roi_data 64 | rois_label_retain = Variable(rois_label.long()) 65 | box_info = Variable(box_info) 66 | rois_label = Variable(rois_label.view(-1).long()) 67 | rois_target = Variable(rois_target.view(-1, rois_target.size(2))) 68 | rois_inside_ws = Variable(rois_inside_ws.view(-1, rois_inside_ws.size(2))) 69 | rois_outside_ws = Variable(rois_outside_ws.view(-1, rois_outside_ws.size(2))) 70 | 71 | else: 72 | rois_label = None 73 | rois_target = None 74 | rois_inside_ws = None 75 | rois_outside_ws = None 76 | rpn_loss_cls = 0 77 | rpn_loss_bbox = 0 78 | 79 | 80 | rois = Variable(rois) 81 | rois_padded = Variable(self.enlarge_bbox(im_info, rois, 0.3)) 82 | 83 | # do roi pooling based on predicted rois 84 | if cfg.POOLING_MODE == 'align': 85 | pooled_feat = self.RCNN_roi_align(base_feat, rois.view(-1, 5)) 86 | pooled_feat_padded = self.RCNN_roi_align(base_feat, rois_padded.view(-1, 5)) 87 | elif cfg.POOLING_MODE == 'pool': 88 | pooled_feat = self.RCNN_roi_pool(base_feat, rois.view(-1,5)) 89 | pooled_feat_padded = self.RCNN_roi_pool(base_feat, rois_padded.view(-1,5)) 90 | 91 | 92 | # feed pooled features to top model 93 | pooled_feat = self._head_to_tail(pooled_feat) 94 | pooled_feat_padded = self._head_to_tail(pooled_feat_padded) 95 | 96 | # compute bbox offset 97 | bbox_pred = self.RCNN_bbox_pred(pooled_feat) 98 | if self.training and not self.class_agnostic: 99 | # select the corresponding columns according to roi labels 100 | bbox_pred_view = bbox_pred.view(bbox_pred.size(0), int(bbox_pred.size(1) / 4), 4) 101 | bbox_pred_select = torch.gather(bbox_pred_view, 1, rois_label.view(rois_label.size(0), 1, 1).expand(rois_label.size(0), 1, 4)) 102 | bbox_pred = bbox_pred_select.squeeze(1) 103 | 104 | # compute object classification probability 105 | cls_score = self.RCNN_cls_score(pooled_feat) 106 | cls_prob = F.softmax(cls_score, 1) 107 | # object_feat = pooled_feat[rois_label==1,:] 108 | # result = self.lineartrial(object_feat) 109 | # extension layer 110 | RCNN_loss_cls = 0 111 | RCNN_loss_bbox = 0 112 | loss_list = [] 113 | 114 | if self.training: 115 | # classification loss 116 | RCNN_loss_cls = F.cross_entropy(cls_score, rois_label) 117 | 118 | # bounding box regression L1 loss 119 | RCNN_loss_bbox = _smooth_l1_loss(bbox_pred, rois_target, rois_inside_ws, rois_outside_ws) 120 | 121 | # auxiliary layer 122 | loss_list = self.extension_layer(pooled_feat, pooled_feat_padded, rois_label_retain, box_info) 123 | else: 124 | loss_list = self.extension_layer(pooled_feat, pooled_feat_padded, None, box_info) 125 | 126 | 127 | 128 | 129 | 130 | cls_prob = cls_prob.view(batch_size, rois.size(1), -1) 131 | bbox_pred = bbox_pred.view(batch_size, rois.size(1), -1) 132 | 133 | return rois, cls_prob, bbox_pred, rpn_loss_cls, rpn_loss_bbox, RCNN_loss_cls, RCNN_loss_bbox, rois_label, loss_list 134 | 135 | def enlarge_bbox(self, im_info, rois, ratio=0.5): 136 | rois_width, rois_height = (rois[:,:,3]-rois[:,:,1]), (rois[:,:,4]-rois[:,:,2]) 137 | rois_padded = rois.clone() 138 | rois_padded[:,:,1] = rois_padded[:,:,1] - ratio*rois_width 139 | rois_padded[:,:,2] = rois_padded[:,:,2] - ratio*rois_height 140 | rois_padded[:,:,1][rois_padded[:,:,1] < 0] = 0 141 | rois_padded[:,:,2][rois_padded[:,:,2] < 0] = 0 142 | 143 | rois_padded[:,:,3] = rois_padded[:,:,3] + ratio*rois_width 144 | rois_padded[:,:,4] = rois_padded[:,:,4] + ratio*rois_height 145 | rois_padded[:,:,3][rois_padded[:,:,3] > im_info[:,0]] = im_info[:,0] 146 | rois_padded[:,:,4][rois_padded[:,:,4] > im_info[:,1]] = im_info[:,1] 147 | return rois_padded 148 | 149 | def _init_weights(self): 150 | def normal_init(m, mean, stddev, truncated=False): 151 | """ 152 | weight initalizer: truncated normal and random normal. 153 | """ 154 | # x is a parameter 155 | if truncated: 156 | m.weight.data.normal_().fmod_(2).mul_(stddev).add_(mean) # not a perfect approximation 157 | else: 158 | m.weight.data.normal_(mean, stddev) 159 | m.bias.data.zero_() 160 | 161 | normal_init(self.RCNN_rpn.RPN_Conv, 0, 0.01, cfg.TRAIN.TRUNCATED) 162 | normal_init(self.RCNN_rpn.RPN_cls_score, 0, 0.01, cfg.TRAIN.TRUNCATED) 163 | normal_init(self.RCNN_rpn.RPN_bbox_pred, 0, 0.01, cfg.TRAIN.TRUNCATED) 164 | normal_init(self.RCNN_cls_score, 0, 0.01, cfg.TRAIN.TRUNCATED) 165 | normal_init(self.RCNN_bbox_pred, 0, 0.001, cfg.TRAIN.TRUNCATED) 166 | 167 | def create_architecture(self): 168 | self._init_modules() 169 | self._init_weights() 170 | -------------------------------------------------------------------------------- /lib/model/roi_align/src/roi_align.c: -------------------------------------------------------------------------------- 1 | #include 2 | #include 3 | #include 4 | 5 | 6 | void ROIAlignForwardCpu(const float* bottom_data, const float spatial_scale, const int num_rois, 7 | const int height, const int width, const int channels, 8 | const int aligned_height, const int aligned_width, const float * bottom_rois, 9 | float* top_data); 10 | 11 | void ROIAlignBackwardCpu(const float* top_diff, const float spatial_scale, const int num_rois, 12 | const int height, const int width, const int channels, 13 | const int aligned_height, const int aligned_width, const float * bottom_rois, 14 | float* top_data); 15 | 16 | int roi_align_forward(int aligned_height, int aligned_width, float spatial_scale, 17 | THFloatTensor * features, THFloatTensor * rois, THFloatTensor * output) 18 | { 19 | //Grab the input tensor 20 | float * data_flat = THFloatTensor_data(features); 21 | float * rois_flat = THFloatTensor_data(rois); 22 | 23 | float * output_flat = THFloatTensor_data(output); 24 | 25 | // Number of ROIs 26 | int num_rois = THFloatTensor_size(rois, 0); 27 | int size_rois = THFloatTensor_size(rois, 1); 28 | if (size_rois != 5) 29 | { 30 | return 0; 31 | } 32 | 33 | // data height 34 | int data_height = THFloatTensor_size(features, 2); 35 | // data width 36 | int data_width = THFloatTensor_size(features, 3); 37 | // Number of channels 38 | int num_channels = THFloatTensor_size(features, 1); 39 | 40 | // do ROIAlignForward 41 | ROIAlignForwardCpu(data_flat, spatial_scale, num_rois, data_height, data_width, num_channels, 42 | aligned_height, aligned_width, rois_flat, output_flat); 43 | 44 | return 1; 45 | } 46 | 47 | int roi_align_backward(int aligned_height, int aligned_width, float spatial_scale, 48 | THFloatTensor * top_grad, THFloatTensor * rois, THFloatTensor * bottom_grad) 49 | { 50 | //Grab the input tensor 51 | float * top_grad_flat = THFloatTensor_data(top_grad); 52 | float * rois_flat = THFloatTensor_data(rois); 53 | 54 | float * bottom_grad_flat = THFloatTensor_data(bottom_grad); 55 | 56 | // Number of ROIs 57 | int num_rois = THFloatTensor_size(rois, 0); 58 | int size_rois = THFloatTensor_size(rois, 1); 59 | if (size_rois != 5) 60 | { 61 | return 0; 62 | } 63 | 64 | // batch size 65 | // int batch_size = THFloatTensor_size(bottom_grad, 0); 66 | // data height 67 | int data_height = THFloatTensor_size(bottom_grad, 2); 68 | // data width 69 | int data_width = THFloatTensor_size(bottom_grad, 3); 70 | // Number of channels 71 | int num_channels = THFloatTensor_size(bottom_grad, 1); 72 | 73 | // do ROIAlignBackward 74 | ROIAlignBackwardCpu(top_grad_flat, spatial_scale, num_rois, data_height, 75 | data_width, num_channels, aligned_height, aligned_width, rois_flat, bottom_grad_flat); 76 | 77 | return 1; 78 | } 79 | 80 | void ROIAlignForwardCpu(const float* bottom_data, const float spatial_scale, const int num_rois, 81 | const int height, const int width, const int channels, 82 | const int aligned_height, const int aligned_width, const float * bottom_rois, 83 | float* top_data) 84 | { 85 | const int output_size = num_rois * aligned_height * aligned_width * channels; 86 | 87 | int idx = 0; 88 | for (idx = 0; idx < output_size; ++idx) 89 | { 90 | // (n, c, ph, pw) is an element in the aligned output 91 | int pw = idx % aligned_width; 92 | int ph = (idx / aligned_width) % aligned_height; 93 | int c = (idx / aligned_width / aligned_height) % channels; 94 | int n = idx / aligned_width / aligned_height / channels; 95 | 96 | float roi_batch_ind = bottom_rois[n * 5 + 0]; 97 | float roi_start_w = bottom_rois[n * 5 + 1] * spatial_scale; 98 | float roi_start_h = bottom_rois[n * 5 + 2] * spatial_scale; 99 | float roi_end_w = bottom_rois[n * 5 + 3] * spatial_scale; 100 | float roi_end_h = bottom_rois[n * 5 + 4] * spatial_scale; 101 | 102 | // Force malformed ROI to be 1x1 103 | float roi_width = fmaxf(roi_end_w - roi_start_w + 1., 0.); 104 | float roi_height = fmaxf(roi_end_h - roi_start_h + 1., 0.); 105 | float bin_size_h = roi_height / (aligned_height - 1.); 106 | float bin_size_w = roi_width / (aligned_width - 1.); 107 | 108 | float h = (float)(ph) * bin_size_h + roi_start_h; 109 | float w = (float)(pw) * bin_size_w + roi_start_w; 110 | 111 | int hstart = fminf(floor(h), height - 2); 112 | int wstart = fminf(floor(w), width - 2); 113 | 114 | int img_start = roi_batch_ind * channels * height * width; 115 | 116 | // bilinear interpolation 117 | if (h < 0 || h >= height || w < 0 || w >= width) 118 | { 119 | top_data[idx] = 0.; 120 | } 121 | else 122 | { 123 | float h_ratio = h - (float)(hstart); 124 | float w_ratio = w - (float)(wstart); 125 | int upleft = img_start + (c * height + hstart) * width + wstart; 126 | int upright = upleft + 1; 127 | int downleft = upleft + width; 128 | int downright = downleft + 1; 129 | 130 | top_data[idx] = bottom_data[upleft] * (1. - h_ratio) * (1. - w_ratio) 131 | + bottom_data[upright] * (1. - h_ratio) * w_ratio 132 | + bottom_data[downleft] * h_ratio * (1. - w_ratio) 133 | + bottom_data[downright] * h_ratio * w_ratio; 134 | } 135 | } 136 | } 137 | 138 | void ROIAlignBackwardCpu(const float* top_diff, const float spatial_scale, const int num_rois, 139 | const int height, const int width, const int channels, 140 | const int aligned_height, const int aligned_width, const float * bottom_rois, 141 | float* bottom_diff) 142 | { 143 | const int output_size = num_rois * aligned_height * aligned_width * channels; 144 | 145 | int idx = 0; 146 | for (idx = 0; idx < output_size; ++idx) 147 | { 148 | // (n, c, ph, pw) is an element in the aligned output 149 | int pw = idx % aligned_width; 150 | int ph = (idx / aligned_width) % aligned_height; 151 | int c = (idx / aligned_width / aligned_height) % channels; 152 | int n = idx / aligned_width / aligned_height / channels; 153 | 154 | float roi_batch_ind = bottom_rois[n * 5 + 0]; 155 | float roi_start_w = bottom_rois[n * 5 + 1] * spatial_scale; 156 | float roi_start_h = bottom_rois[n * 5 + 2] * spatial_scale; 157 | float roi_end_w = bottom_rois[n * 5 + 3] * spatial_scale; 158 | float roi_end_h = bottom_rois[n * 5 + 4] * spatial_scale; 159 | 160 | // Force malformed ROI to be 1x1 161 | float roi_width = fmaxf(roi_end_w - roi_start_w + 1., 0.); 162 | float roi_height = fmaxf(roi_end_h - roi_start_h + 1., 0.); 163 | float bin_size_h = roi_height / (aligned_height - 1.); 164 | float bin_size_w = roi_width / (aligned_width - 1.); 165 | 166 | float h = (float)(ph) * bin_size_h + roi_start_h; 167 | float w = (float)(pw) * bin_size_w + roi_start_w; 168 | 169 | int hstart = fminf(floor(h), height - 2); 170 | int wstart = fminf(floor(w), width - 2); 171 | 172 | int img_start = roi_batch_ind * channels * height * width; 173 | 174 | // bilinear interpolation 175 | if (h < 0 || h >= height || w < 0 || w >= width) 176 | { 177 | float h_ratio = h - (float)(hstart); 178 | float w_ratio = w - (float)(wstart); 179 | int upleft = img_start + (c * height + hstart) * width + wstart; 180 | int upright = upleft + 1; 181 | int downleft = upleft + width; 182 | int downright = downleft + 1; 183 | 184 | bottom_diff[upleft] += top_diff[idx] * (1. - h_ratio) * (1. - w_ratio); 185 | bottom_diff[upright] += top_diff[idx] * (1. - h_ratio) * w_ratio; 186 | bottom_diff[downleft] += top_diff[idx] * h_ratio * (1. - w_ratio); 187 | bottom_diff[downright] += top_diff[idx] * h_ratio * w_ratio; 188 | } 189 | } 190 | } 191 | -------------------------------------------------------------------------------- /lib/model/roi_align/src/roi_align_kernel.cu: -------------------------------------------------------------------------------- 1 | #ifdef __cplusplus 2 | extern "C" { 3 | #endif 4 | 5 | #include 6 | #include 7 | #include 8 | #include "roi_align_kernel.h" 9 | 10 | #define CUDA_1D_KERNEL_LOOP(i, n) \ 11 | for (int i = blockIdx.x * blockDim.x + threadIdx.x; i < n; \ 12 | i += blockDim.x * gridDim.x) 13 | 14 | 15 | __global__ void ROIAlignForward(const int nthreads, const float* bottom_data, const float spatial_scale, const int height, const int width, 16 | const int channels, const int aligned_height, const int aligned_width, const float* bottom_rois, float* top_data) { 17 | CUDA_1D_KERNEL_LOOP(index, nthreads) { 18 | // (n, c, ph, pw) is an element in the aligned output 19 | // int n = index; 20 | // int pw = n % aligned_width; 21 | // n /= aligned_width; 22 | // int ph = n % aligned_height; 23 | // n /= aligned_height; 24 | // int c = n % channels; 25 | // n /= channels; 26 | 27 | int pw = index % aligned_width; 28 | int ph = (index / aligned_width) % aligned_height; 29 | int c = (index / aligned_width / aligned_height) % channels; 30 | int n = index / aligned_width / aligned_height / channels; 31 | 32 | // bottom_rois += n * 5; 33 | float roi_batch_ind = bottom_rois[n * 5 + 0]; 34 | float roi_start_w = bottom_rois[n * 5 + 1] * spatial_scale; 35 | float roi_start_h = bottom_rois[n * 5 + 2] * spatial_scale; 36 | float roi_end_w = bottom_rois[n * 5 + 3] * spatial_scale; 37 | float roi_end_h = bottom_rois[n * 5 + 4] * spatial_scale; 38 | 39 | // Force malformed ROIs to be 1x1 40 | float roi_width = fmaxf(roi_end_w - roi_start_w + 1., 0.); 41 | float roi_height = fmaxf(roi_end_h - roi_start_h + 1., 0.); 42 | float bin_size_h = roi_height / (aligned_height - 1.); 43 | float bin_size_w = roi_width / (aligned_width - 1.); 44 | 45 | float h = (float)(ph) * bin_size_h + roi_start_h; 46 | float w = (float)(pw) * bin_size_w + roi_start_w; 47 | 48 | int hstart = fminf(floor(h), height - 2); 49 | int wstart = fminf(floor(w), width - 2); 50 | 51 | int img_start = roi_batch_ind * channels * height * width; 52 | 53 | // bilinear interpolation 54 | if (h < 0 || h >= height || w < 0 || w >= width) { 55 | top_data[index] = 0.; 56 | } else { 57 | float h_ratio = h - (float)(hstart); 58 | float w_ratio = w - (float)(wstart); 59 | int upleft = img_start + (c * height + hstart) * width + wstart; 60 | int upright = upleft + 1; 61 | int downleft = upleft + width; 62 | int downright = downleft + 1; 63 | 64 | top_data[index] = bottom_data[upleft] * (1. - h_ratio) * (1. - w_ratio) 65 | + bottom_data[upright] * (1. - h_ratio) * w_ratio 66 | + bottom_data[downleft] * h_ratio * (1. - w_ratio) 67 | + bottom_data[downright] * h_ratio * w_ratio; 68 | } 69 | } 70 | } 71 | 72 | 73 | int ROIAlignForwardLaucher(const float* bottom_data, const float spatial_scale, const int num_rois, const int height, const int width, 74 | const int channels, const int aligned_height, const int aligned_width, const float* bottom_rois, float* top_data, cudaStream_t stream) { 75 | const int kThreadsPerBlock = 1024; 76 | const int output_size = num_rois * aligned_height * aligned_width * channels; 77 | cudaError_t err; 78 | 79 | 80 | ROIAlignForward<<<(output_size + kThreadsPerBlock - 1) / kThreadsPerBlock, kThreadsPerBlock, 0, stream>>>( 81 | output_size, bottom_data, spatial_scale, height, width, channels, 82 | aligned_height, aligned_width, bottom_rois, top_data); 83 | 84 | err = cudaGetLastError(); 85 | if(cudaSuccess != err) { 86 | fprintf( stderr, "cudaCheckError() failed : %s\n", cudaGetErrorString( err ) ); 87 | exit( -1 ); 88 | } 89 | 90 | return 1; 91 | } 92 | 93 | 94 | __global__ void ROIAlignBackward(const int nthreads, const float* top_diff, const float spatial_scale, const int height, const int width, 95 | const int channels, const int aligned_height, const int aligned_width, float* bottom_diff, const float* bottom_rois) { 96 | CUDA_1D_KERNEL_LOOP(index, nthreads) { 97 | 98 | // (n, c, ph, pw) is an element in the aligned output 99 | int pw = index % aligned_width; 100 | int ph = (index / aligned_width) % aligned_height; 101 | int c = (index / aligned_width / aligned_height) % channels; 102 | int n = index / aligned_width / aligned_height / channels; 103 | 104 | float roi_batch_ind = bottom_rois[n * 5 + 0]; 105 | float roi_start_w = bottom_rois[n * 5 + 1] * spatial_scale; 106 | float roi_start_h = bottom_rois[n * 5 + 2] * spatial_scale; 107 | float roi_end_w = bottom_rois[n * 5 + 3] * spatial_scale; 108 | float roi_end_h = bottom_rois[n * 5 + 4] * spatial_scale; 109 | /* int roi_start_w = round(bottom_rois[1] * spatial_scale); */ 110 | /* int roi_start_h = round(bottom_rois[2] * spatial_scale); */ 111 | /* int roi_end_w = round(bottom_rois[3] * spatial_scale); */ 112 | /* int roi_end_h = round(bottom_rois[4] * spatial_scale); */ 113 | 114 | // Force malformed ROIs to be 1x1 115 | float roi_width = fmaxf(roi_end_w - roi_start_w + 1., 0.); 116 | float roi_height = fmaxf(roi_end_h - roi_start_h + 1., 0.); 117 | float bin_size_h = roi_height / (aligned_height - 1.); 118 | float bin_size_w = roi_width / (aligned_width - 1.); 119 | 120 | float h = (float)(ph) * bin_size_h + roi_start_h; 121 | float w = (float)(pw) * bin_size_w + roi_start_w; 122 | 123 | int hstart = fminf(floor(h), height - 2); 124 | int wstart = fminf(floor(w), width - 2); 125 | 126 | int img_start = roi_batch_ind * channels * height * width; 127 | 128 | // bilinear interpolation 129 | if (!(h < 0 || h >= height || w < 0 || w >= width)) { 130 | float h_ratio = h - (float)(hstart); 131 | float w_ratio = w - (float)(wstart); 132 | int upleft = img_start + (c * height + hstart) * width + wstart; 133 | int upright = upleft + 1; 134 | int downleft = upleft + width; 135 | int downright = downleft + 1; 136 | 137 | atomicAdd(bottom_diff + upleft, top_diff[index] * (1. - h_ratio) * (1 - w_ratio)); 138 | atomicAdd(bottom_diff + upright, top_diff[index] * (1. - h_ratio) * w_ratio); 139 | atomicAdd(bottom_diff + downleft, top_diff[index] * h_ratio * (1 - w_ratio)); 140 | atomicAdd(bottom_diff + downright, top_diff[index] * h_ratio * w_ratio); 141 | } 142 | } 143 | } 144 | 145 | int ROIAlignBackwardLaucher(const float* top_diff, const float spatial_scale, const int batch_size, const int num_rois, const int height, const int width, 146 | const int channels, const int aligned_height, const int aligned_width, const float* bottom_rois, float* bottom_diff, cudaStream_t stream) { 147 | const int kThreadsPerBlock = 1024; 148 | const int output_size = num_rois * aligned_height * aligned_width * channels; 149 | cudaError_t err; 150 | 151 | ROIAlignBackward<<<(output_size + kThreadsPerBlock - 1) / kThreadsPerBlock, kThreadsPerBlock, 0, stream>>>( 152 | output_size, top_diff, spatial_scale, height, width, channels, 153 | aligned_height, aligned_width, bottom_diff, bottom_rois); 154 | 155 | err = cudaGetLastError(); 156 | if(cudaSuccess != err) { 157 | fprintf( stderr, "cudaCheckError() failed : %s\n", cudaGetErrorString( err ) ); 158 | exit( -1 ); 159 | } 160 | 161 | return 1; 162 | } 163 | 164 | 165 | #ifdef __cplusplus 166 | } 167 | #endif 168 | -------------------------------------------------------------------------------- /lib/pycocotools/maskApi.c: -------------------------------------------------------------------------------- 1 | /************************************************************************** 2 | * Microsoft COCO Toolbox. version 2.0 3 | * Data, paper, and tutorials available at: http://mscoco.org/ 4 | * Code written by Piotr Dollar and Tsung-Yi Lin, 2015. 5 | * Licensed under the Simplified BSD License [see coco/license.txt] 6 | **************************************************************************/ 7 | #include "maskApi.h" 8 | #include 9 | #include 10 | 11 | uint umin( uint a, uint b ) { return (ab) ? a : b; } 13 | 14 | void rleInit( RLE *R, siz h, siz w, siz m, uint *cnts ) { 15 | R->h=h; R->w=w; R->m=m; R->cnts=(m==0)?0:malloc(sizeof(uint)*m); 16 | if(cnts) for(siz j=0; jcnts[j]=cnts[j]; 17 | } 18 | 19 | void rleFree( RLE *R ) { 20 | free(R->cnts); R->cnts=0; 21 | } 22 | 23 | void rlesInit( RLE **R, siz n ) { 24 | *R = (RLE*) malloc(sizeof(RLE)*n); 25 | for(siz i=0; i0 ) { 61 | c=umin(ca,cb); cc+=c; ct=0; 62 | ca-=c; if(!ca && a0) { 83 | crowd=iscrowd!=NULL && iscrowd[g]; 84 | if(dt[d].h!=gt[g].h || dt[d].w!=gt[g].w) { o[g*m+d]=-1; continue; } 85 | siz ka, kb, a, b; uint c, ca, cb, ct, i, u; bool va, vb; 86 | ca=dt[d].cnts[0]; ka=dt[d].m; va=vb=0; 87 | cb=gt[g].cnts[0]; kb=gt[g].m; a=b=1; i=u=0; ct=1; 88 | while( ct>0 ) { 89 | c=umin(ca,cb); if(va||vb) { u+=c; if(va&&vb) i+=c; } ct=0; 90 | ca-=c; if(!ca && ad?1:c=dy && xs>xe) || (dxye); 151 | if(flip) { t=xs; xs=xe; xe=t; t=ys; ys=ye; ye=t; } 152 | s = dx>=dy ? (double)(ye-ys)/dx : (double)(xe-xs)/dy; 153 | if(dx>=dy) for( int d=0; d<=dx; d++ ) { 154 | t=flip?dx-d:d; u[m]=t+xs; v[m]=(int)(ys+s*t+.5); m++; 155 | } else for( int d=0; d<=dy; d++ ) { 156 | t=flip?dy-d:d; v[m]=t+ys; u[m]=(int)(xs+s*t+.5); m++; 157 | } 158 | } 159 | // get points along y-boundary and downsample 160 | free(x); free(y); k=m; m=0; double xd, yd; 161 | x=malloc(sizeof(int)*k); y=malloc(sizeof(int)*k); 162 | for( j=1; jw-1 ) continue; 165 | yd=(double)(v[j]h) yd=h; yd=ceil(yd); 167 | x[m]=(int) xd; y[m]=(int) yd; m++; 168 | } 169 | // compute rle encoding given y-boundary points 170 | k=m; a=malloc(sizeof(uint)*(k+1)); 171 | for( j=0; j0) b[m++]=a[j++]; else { 177 | j++; if(jm, p=0; long x; bool more; 184 | char *s=malloc(sizeof(char)*m*6); 185 | for( i=0; icnts[i]; if(i>2) x-=(long) R->cnts[i-2]; more=1; 187 | while( more ) { 188 | char c=x & 0x1f; x >>= 5; more=(c & 0x10) ? x!=-1 : x!=0; 189 | if(more) c |= 0x20; c+=48; s[p++]=c; 190 | } 191 | } 192 | s[p]=0; return s; 193 | } 194 | 195 | void rleFrString( RLE *R, char *s, siz h, siz w ) { 196 | siz m=0, p=0, k; long x; bool more; uint *cnts; 197 | while( s[m] ) m++; cnts=malloc(sizeof(uint)*m); m=0; 198 | while( s[p] ) { 199 | x=0; k=0; more=1; 200 | while( more ) { 201 | char c=s[p]-48; x |= (c & 0x1f) << 5*k; 202 | more = c & 0x20; p++; k++; 203 | if(!more && (c & 0x10)) x |= -1 << 5*k; 204 | } 205 | if(m>2) x+=(long) cnts[m-2]; cnts[m++]=(uint) x; 206 | } 207 | rleInit(R,h,w,m,cnts); free(cnts); 208 | } 209 | -------------------------------------------------------------------------------- /lib/model/csrc/cuda/ROIPool_cuda.cu: -------------------------------------------------------------------------------- 1 | // Copyright (c) Facebook, Inc. and its affiliates. All Rights Reserved. 2 | #include 3 | #include 4 | #include 5 | 6 | // #include 7 | // #include 8 | // #include 9 | 10 | // template 11 | // __host__ __device__ __forceinline__ T THCCeilDiv(T a, T b) { 12 | // return (a + b - 1) / b; 13 | // } 14 | // TODO make it in a common file 15 | #define CUDA_1D_KERNEL_LOOP(i, n) \ 16 | for (int i = blockIdx.x * blockDim.x + threadIdx.x; i < n; \ 17 | i += blockDim.x * gridDim.x) 18 | 19 | 20 | template 21 | __global__ void RoIPoolFForward(const int nthreads, const T* bottom_data, 22 | const T spatial_scale, const int channels, const int height, 23 | const int width, const int pooled_height, const int pooled_width, 24 | const T* bottom_rois, T* top_data, int* argmax_data) { 25 | CUDA_1D_KERNEL_LOOP(index, nthreads) { 26 | // (n, c, ph, pw) is an element in the pooled output 27 | int pw = index % pooled_width; 28 | int ph = (index / pooled_width) % pooled_height; 29 | int c = (index / pooled_width / pooled_height) % channels; 30 | int n = index / pooled_width / pooled_height / channels; 31 | 32 | const T* offset_bottom_rois = bottom_rois + n * 5; 33 | int roi_batch_ind = offset_bottom_rois[0]; 34 | int roi_start_w = round(offset_bottom_rois[1] * spatial_scale); 35 | int roi_start_h = round(offset_bottom_rois[2] * spatial_scale); 36 | int roi_end_w = round(offset_bottom_rois[3] * spatial_scale); 37 | int roi_end_h = round(offset_bottom_rois[4] * spatial_scale); 38 | 39 | // Force malformed ROIs to be 1x1 40 | int roi_width = max(roi_end_w - roi_start_w + 1, 1); 41 | int roi_height = max(roi_end_h - roi_start_h + 1, 1); 42 | T bin_size_h = static_cast(roi_height) 43 | / static_cast(pooled_height); 44 | T bin_size_w = static_cast(roi_width) 45 | / static_cast(pooled_width); 46 | 47 | int hstart = static_cast(floor(static_cast(ph) 48 | * bin_size_h)); 49 | int wstart = static_cast(floor(static_cast(pw) 50 | * bin_size_w)); 51 | int hend = static_cast(ceil(static_cast(ph + 1) 52 | * bin_size_h)); 53 | int wend = static_cast(ceil(static_cast(pw + 1) 54 | * bin_size_w)); 55 | 56 | // Add roi offsets and clip to input boundaries 57 | hstart = min(max(hstart + roi_start_h, 0), height); 58 | hend = min(max(hend + roi_start_h, 0), height); 59 | wstart = min(max(wstart + roi_start_w, 0), width); 60 | wend = min(max(wend + roi_start_w, 0), width); 61 | bool is_empty = (hend <= hstart) || (wend <= wstart); 62 | 63 | // Define an empty pooling region to be zero 64 | T maxval = is_empty ? 0 : -FLT_MAX; 65 | // If nothing is pooled, argmax = -1 causes nothing to be backprop'd 66 | int maxidx = -1; 67 | const T* offset_bottom_data = 68 | bottom_data + (roi_batch_ind * channels + c) * height * width; 69 | for (int h = hstart; h < hend; ++h) { 70 | for (int w = wstart; w < wend; ++w) { 71 | int bottom_index = h * width + w; 72 | if (offset_bottom_data[bottom_index] > maxval) { 73 | maxval = offset_bottom_data[bottom_index]; 74 | maxidx = bottom_index; 75 | } 76 | } 77 | } 78 | top_data[index] = maxval; 79 | argmax_data[index] = maxidx; 80 | } 81 | } 82 | 83 | template 84 | __global__ void RoIPoolFBackward(const int nthreads, const T* top_diff, 85 | const int* argmax_data, const int num_rois, const T spatial_scale, 86 | const int channels, const int height, const int width, 87 | const int pooled_height, const int pooled_width, T* bottom_diff, 88 | const T* bottom_rois) { 89 | CUDA_1D_KERNEL_LOOP(index, nthreads) { 90 | // (n, c, ph, pw) is an element in the pooled output 91 | int pw = index % pooled_width; 92 | int ph = (index / pooled_width) % pooled_height; 93 | int c = (index / pooled_width / pooled_height) % channels; 94 | int n = index / pooled_width / pooled_height / channels; 95 | 96 | const T* offset_bottom_rois = bottom_rois + n * 5; 97 | int roi_batch_ind = offset_bottom_rois[0]; 98 | int bottom_offset = (roi_batch_ind * channels + c) * height * width; 99 | int top_offset = (n * channels + c) * pooled_height * pooled_width; 100 | const T* offset_top_diff = top_diff + top_offset; 101 | T* offset_bottom_diff = bottom_diff + bottom_offset; 102 | const int* offset_argmax_data = argmax_data + top_offset; 103 | 104 | int argmax = offset_argmax_data[ph * pooled_width + pw]; 105 | if (argmax != -1) { 106 | atomicAdd( 107 | offset_bottom_diff + argmax, 108 | static_cast(offset_top_diff[ph * pooled_width + pw])); 109 | 110 | } 111 | } 112 | } 113 | 114 | std::tuple ROIPool_forward_cuda(const at::Tensor& input, 115 | const at::Tensor& rois, 116 | const float spatial_scale, 117 | const int pooled_height, 118 | const int pooled_width) { 119 | AT_ASSERTM(input.type().is_cuda(), "input must be a CUDA tensor"); 120 | AT_ASSERTM(rois.type().is_cuda(), "rois must be a CUDA tensor"); 121 | 122 | auto num_rois = rois.size(0); 123 | auto channels = input.size(1); 124 | auto height = input.size(2); 125 | auto width = input.size(3); 126 | 127 | auto output = at::empty({num_rois, channels, pooled_height, pooled_width}, input.options()); 128 | auto output_size = num_rois * pooled_height * pooled_width * channels; 129 | auto argmax = at::zeros({num_rois, channels, pooled_height, pooled_width}, input.options().dtype(at::kInt)); 130 | 131 | cudaStream_t stream = at::cuda::getCurrentCUDAStream(); 132 | 133 | dim3 grid(std::min(at::ceil_div(output_size, 512L), 4096L)); 134 | dim3 block(512); 135 | 136 | if (output.numel() == 0) { 137 | C10_CUDA_CHECK(cudaGetLastError()); 138 | return std::make_tuple(output, argmax); 139 | } 140 | 141 | AT_DISPATCH_FLOATING_TYPES(input.type(), "ROIPool_forward", [&] { 142 | RoIPoolFForward<<>>( 143 | output_size, 144 | input.contiguous().data(), 145 | spatial_scale, 146 | channels, 147 | height, 148 | width, 149 | pooled_height, 150 | pooled_width, 151 | rois.contiguous().data(), 152 | output.data(), 153 | argmax.data()); 154 | }); 155 | C10_CUDA_CHECK(cudaGetLastError()); 156 | return std::make_tuple(output, argmax); 157 | } 158 | 159 | // TODO remove the dependency on input and use instead its sizes -> save memory 160 | at::Tensor ROIPool_backward_cuda(const at::Tensor& grad, 161 | const at::Tensor& input, 162 | const at::Tensor& rois, 163 | const at::Tensor& argmax, 164 | const float spatial_scale, 165 | const int pooled_height, 166 | const int pooled_width, 167 | const int batch_size, 168 | const int channels, 169 | const int height, 170 | const int width) { 171 | AT_ASSERTM(grad.type().is_cuda(), "grad must be a CUDA tensor"); 172 | AT_ASSERTM(rois.type().is_cuda(), "rois must be a CUDA tensor"); 173 | // TODO add more checks 174 | 175 | auto num_rois = rois.size(0); 176 | auto grad_input = at::zeros({batch_size, channels, height, width}, grad.options()); 177 | 178 | cudaStream_t stream = at::cuda::getCurrentCUDAStream(); 179 | 180 | dim3 grid(std::min(at::ceil_div(grad.numel(), 512L), 4096L)); 181 | dim3 block(512); 182 | 183 | // handle possibly empty gradients 184 | if (grad.numel() == 0) { 185 | C10_CUDA_CHECK(cudaGetLastError()); 186 | return grad_input; 187 | } 188 | 189 | AT_DISPATCH_FLOATING_TYPES(grad.type(), "ROIPool_backward", [&] { 190 | RoIPoolFBackward<<>>( 191 | grad.numel(), 192 | grad.contiguous().data(), 193 | argmax.data(), 194 | num_rois, 195 | spatial_scale, 196 | channels, 197 | height, 198 | width, 199 | pooled_height, 200 | pooled_width, 201 | grad_input.data(), 202 | rois.contiguous().data()); 203 | }); 204 | C10_CUDA_CHECK(cudaGetLastError()); 205 | return grad_input; 206 | } --------------------------------------------------------------------------------