├── Date.md
└── README.md


/Date.md:
--------------------------------------------------------------------------------
  1 | 2019-10-11 Update 1 Project
  2 | 
  3 | **Detectron2**
  4 | 
  5 | - intro: Detectron2 is FAIR's next-generation research platform for object detection and segmentation.
  6 | - [blog](https://ai.facebook.com/blog/-detectron2-a-pytorch-based-modular-object-detection-library-/)
  7 | - code: <https://github.com/facebookresearch/detectron2>
  8 | 
  9 | 2019-09-06 Update 1 paper
 10 | 
 11 | **Imbalance Problems in Object Detection: A Review**
 12 | 
 13 | - intro: under review at TPAMI
 14 | - arXiv: <https://arxiv.org/abs/1909.00169>
 15 | 
 16 | 2019-08-14 Update 1 paper
 17 | 
 18 | **Recent Advances in Deep Learning for Object Detection**
 19 | 
 20 | - intro: From 2013 (OverFeat) to 2019 (DetNAS)
 21 | - arXiv: <https://arxiv.org/abs/1908.03673>
 22 | 
 23 | 2019-07-24 Update 1 paper
 24 | 
 25 | **A Survey of Deep Learning-based Object Detection**
 26 | 
 27 | - intro：From Fast R-CNN to NAS-FPN
 28 | 
 29 | - arXiv：<https://arxiv.org/abs/1907.09408>
 30 | 
 31 | 2019-05-17 Update 1 paper
 32 | 
 33 | **Object Detection in 20 Years: A Survey**
 34 | 
 35 | - intro：This work has been submitted to the IEEE TPAMI for possible publication
 36 | - arXiv：<https://arxiv.org/abs/1905.05055>
 37 | 
 38 | 2019-04-05 Update 1 paper
 39 | 
 40 | **Comparison Network for One-Shot Conditional Object Detection**
 41 | 
 42 | - arXiv: https://arxiv.org/abs/1904.02317
 43 | 
 44 | 2019-03-05 Update 1 paper
 45 | 
 46 | **Feature Selective Anchor-Free Module for Single-Shot Object Detection**
 47 | 
 48 | - intro: CVPR 2019
 49 | - arXiv: https://arxiv.org/abs/1903.00621
 50 | 
 51 | 2019-02-15 Update 3 detection toolbox
 52 | 
 53 | - [Detectron(FAIR)](https://github.com/facebookresearch/Detectron): Detectron is Facebook AI Research's software system that implements state-of-the-art object detection algorithms, including [Mask R-CNN](https://arxiv.org/abs/1703.06870). It is written in Python and powered by the [Caffe2](https://github.com/caffe2/caffe2) deep learning framework.
 54 | 
 55 | - [maskrcnn-benchmark(FAIR)](https://github.com/facebookresearch/maskrcnn-benchmark): Fast, modular reference implementation of Instance Segmentation and Object Detection algorithms in PyTorch.
 56 | 
 57 | - [mmdetection(SenseTime&CUHK)](https://github.com/open-mmlab/mmdetection): mmdetection is an open source object detection toolbox based on PyTorch. It is a part of the open-mmlab project developed by [Multimedia Laboratory, CUHK](http://mmlab.ie.cuhk.edu.hk/).
 58 | 
 59 | 2019-01-25 Update 5 papers
 60 | 
 61 | **3D Backbone Network for 3D Object Detection**
 62 | 
 63 | - arXiv: https://arxiv.org/abs/1901.08373
 64 | 
 65 | **Object Detection based on Region Decomposition and Assembly**
 66 | 
 67 | - intro: AAAI 2019
 68 | 
 69 | - arXiv: https://arxiv.org/abs/1901.08225
 70 | 
 71 | **Bottom-up Object Detection by Grouping Extreme and Center Points**
 72 | 
 73 | - intro: one stage 43.2% on COCO test-dev
 74 | - arXiv: https://arxiv.org/abs/1901.08043
 75 | - github: https://github.com/xingyizhou/ExtremeNet
 76 | 
 77 | **ORSIm Detector: A Novel Object Detection Framework in Optical Remote Sensing Imagery Using Spatial-Frequency Channel Features**
 78 | 
 79 | - intro: IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
 80 | 
 81 | - arXiv: https://arxiv.org/abs/1901.07925
 82 | 
 83 | **Consistent Optimization for Single-Shot Object Detection**
 84 | 
 85 | - intro: improves RetinaNet from 39.1 AP to 40.1 AP on COCO datase
 86 | 
 87 | - arXiv: https://arxiv.org/abs/1901.06563
 88 | 
 89 | 2019-01-15 Update 1 paper
 90 | 
 91 | **Learning Pairwise Relationship for Multi-object Detection in Crowded Scenes**
 92 | 
 93 | - arXiv: https://arxiv.org/abs/1901.03796
 94 | 
 95 | 2019-01-14 Update 1 paper
 96 | 
 97 | **RetinaMask: Learning to predict masks improves state-of-the-art single-shot detection for free**
 98 | 
 99 | - arXiv: https://arxiv.org/abs/1901.03353
100 | - github: https://github.com/chengyangfu/retinamask
101 | 
102 | 2019-01-12 Update 1 paper
103 | 
104 | **Region Proposal by Guided Anchoring**
105 | 
106 | - intro: CUHK - SenseTime Joint Lab
107 | - arXiv: https://arxiv.org/abs/1901.03278
108 | 
109 | 2019-01-08 Update 1 paper
110 | 
111 | **Scale-Aware Trident Networks for Object Detection**
112 | 
113 | - intro: mAP of **48.4** on the COCO dataset
114 | - arXiv: https://arxiv.org/abs/1901.01892
115 | 
116 | 2019-01-04 Update 1 paper
117 | 
118 | **Large-Scale Object Detection of Images from Network Cameras in Variable Ambient Lighting Conditions**
119 | 
120 | - arXiv: https://arxiv.org/abs/1812.11901
121 | 
122 | 2018-12-13 Update 1 paper
123 | 
124 | **Strong-Weak Distribution Alignment for Adaptive Object Detection**
125 | 
126 | - arXiv: https://arxiv.org/abs/1812.04798
127 | 
128 | 2018-12-05 Update 3 papers
129 | 
130 | **AutoFocus: Efficient Multi-Scale Inference**
131 | 
132 | - intro: AutoFocus obtains an **mAP of 47.9%** (68.3% at 50% overlap) on the **COCO test-dev** set while processing **6.4 images per second on a Titan X (Pascal) GPU** 
133 | - arXiv: https://arxiv.org/abs/1812.01600
134 | 
135 | **NOTE-RCNN: NOise Tolerant Ensemble RCNN for Semi-Supervised Object Detection**
136 | 
137 | - intro: Google Could
138 | - arXiv: https://arxiv.org/abs/1812.00124
139 | 
140 | **SPLAT: Semantic Pixel-Level Adaptation Transforms for Detection**
141 | 
142 | - intro: UC Berkeley
143 | - arXiv: https://arxiv.org/abs/1812.00929
144 | 
145 | 2018-12-04 Update 10 papers
146 | 
147 | **Grid R-CNN**
148 | 
149 | - intro: SenseTime
150 | - arXiv: https://arxiv.org/abs/1811.12030
151 | 
152 | **Deformable ConvNets v2: More Deformable, Better Results**
153 | 
154 | - intro: Microsoft Research Asia
155 | 
156 | - arXiv: https://arxiv.org/abs/1811.11168
157 | 
158 | **Anchor Box Optimization for Object Detection**
159 | 
160 | - intro: Microsoft Research
161 | - arXiv: https://arxiv.org/abs/1812.00469
162 | 
163 | **Efficient Coarse-to-Fine Non-Local Module for the Detection of Small Objects**
164 | 
165 | - intro: https://arxiv.org/abs/1811.12152
166 | 
167 | **NOTE-RCNN: NOise Tolerant Ensemble RCNN for Semi-Supervised Object Detection**
168 | 
169 | - arXiv: https://arxiv.org/abs/1812.00124
170 | 
171 | **Learning RoI Transformer for Detecting Oriented Objects in Aerial Images**
172 | 
173 | - arXiv: https://arxiv.org/abs/1812.00155
174 | 
175 | **Integrated Object Detection and Tracking with Tracklet-Conditioned Detection**
176 | 
177 | - intro: Microsoft Research Asia
178 | - arXiv: https://arxiv.org/abs/1811.11167
179 | 
180 | **Deep Regionlets: Blended Representation and Deep Learning for Generic Object Detection**
181 | 
182 | - arXiv: https://arxiv.org/abs/1811.11318
183 | 
184 |  **Gradient Harmonized Single-stage Detector**
185 | 
186 | - intro: AAAI 2019
187 | - arXiv: https://arxiv.org/abs/1811.05181
188 | 
189 | **CFENet: Object Detection with Comprehensive Feature Enhancement Module**
190 | 
191 | - intro: ACCV 2018
192 | - github: https://github.com/qijiezhao/CFENet
193 | 
194 | 2018-11-19
195 | 
196 | **DeRPN: Taking a further step toward more general object detection**
197 | 
198 | - intro: AAAI 2019
199 | - arXiv: https://arxiv.org/abs/1811.06700
200 | - github: https://github.com/HCIILAB/DeRPN
201 | 
202 | 2018-11-14
203 | 
204 | **M2Det: A Single-Shot Object Detector based on Multi-Level Feature Pyramid Network**
205 | 
206 | - intro: AAAI 2019
207 | - arXiv: https://arxiv.org/abs/1811.04533
208 | - github: https://github.com/qijiezhao/M2Det
209 | 
210 | 2018-10-31
211 | 
212 | **Hybrid Knowledge Routed Modules for Large-scale Object Detection**
213 | 
214 | - intro: Sun Yat-Sen University & Huawei Noah’s Ark Lab
215 | - arXiv: https://arxiv.org/abs/1810.12681
216 | - github: https://github.com/chanyn/HKRM
217 | 
218 | 2018-10-08
219 | 
220 | **Weakly Supervised Object Detection in Artworks**
221 | 
222 | - intro: ECCV 2018 Workshop Computer Vision for Art Analysis
223 | - arXiv: https://arxiv.org/abs/1810.02569
224 | - Datasets: https://wsoda.telecom-paristech.fr/downloads/dataset/IconArt_v1.zip
225 | 
226 | **Cross-Domain Weakly-Supervised Object Detection through Progressive Domain Adaptation**
227 | 
228 | - intro: CVPR 2018
229 | - arXiv: https://arxiv.org/abs/1803.11365
230 | - homepage: https://naoto0804.github.io/cross_domain_detection/
231 | - paper: http://openaccess.thecvf.com/content_cvpr_2018/html/Inoue_Cross-Domain_Weakly-Supervised_Object_CVPR_2018_paper.html
232 | - github: https://github.com/naoto0804/cross-domain-detection
233 | 
234 | 2018-09-26
235 | 
236 | **Object Detection from Scratch with Deep Supervision**
237 | 
238 | - intro: This is an extended version of DSOD
239 | - arXiv: https://arxiv.org/abs/1809.09294
240 | 
241 | 2018-09-25
242 | 
243 | **《Softer-NMS: Rethinking Bounding Box Regression for Accurate Object Detection》**
244 | 
245 | - intro: CMU & Face++
246 | - arXiv: https://arxiv.org/abs/1809.08545
247 | - github: https://github.com/yihui-he/softer-NMS
248 | 
249 | 2018-09-21
250 | 
251 | **《Receptive Field Block Net for Accurate and Fast Object Detection》**
252 | 
253 | - intro: ECCV 2018
254 | - arXiv: [https://arxiv.org/abs/1711.07767](https://arxiv.org/abs/1711.07767)
255 | - github: [https://github.com/ruinmessi/RFBNet](https://github.com/ruinmessi/RFBNet)
256 | 
257 | 2018-09-11
258 | 
259 | **《Recent Advances in Object Detection in the Age of Deep Convolutional Neural Networks》**
260 | 
261 | - intro: awesome
262 | 
263 | 
264 | - arXiv: https://arxiv.org/abs/1809.03193
265 | 
266 | 2018-09-10
267 | 
268 | **《Deep Learning for Generic Object Detection: A Survey》**
269 | 
270 | - intro: Submitted to IJCV 2018
271 | - arXiv: https://arxiv.org/abs/1809.02165
272 | 
273 | 2018-08-27
274 | 
275 | **Deep Feature Pyramid Reconfiguration for Object Detection**
276 | 
277 | - intro: ECCV 2018
278 | - arXiv: https://arxiv.org/abs/1808.07993
279 | 
280 | 2018-08-17
281 | 
282 | **R3-Net: A Deep Network for Multi-oriented Vehicle Detection in Aerial Images and Videos**
283 | 
284 | - arxiv: https://arxiv.org/abs/1808.05560
285 | - youtube: https://youtu.be/xCYD-tYudN0
286 | 
287 | 2018-08-14
288 | 
289 | **《Unsupervised Hard Example Mining from Videos for Improved Object Detection》**
290 | 
291 | - intro: ECCV 2018
292 | - arXiv: https://arxiv.org/abs/1808.04285
293 | 
294 | 2018-08-10
295 | 
296 | **CornerNet: Detecting Objects as Paired Keypoints**
297 | 
298 | - intro: ECCV 2018
299 | - arXiv: https://arxiv.org/abs/1808.01244
300 | 
301 | 2018-07-30
302 | 
303 | **Acquisition of Localization Confidence for Accurate Object Detection**
304 | 
305 | - intro: ECCV 2018
306 | - arXiv: https://arxiv.org/abs/1807.11590
307 | - github: https://github.com/vacancy/PreciseRoIPooling


--------------------------------------------------------------------------------
/README.md:
--------------------------------------------------------------------------------
   1 | # object-detection
   2 | 
   3 | [TOC]
   4 | 
   5 | This is a list of awesome articles about object detection. If you want to read the paper according to time, you can refer to [Date](Date.md).
   6 | 
   7 | - R-CNN
   8 | - Fast R-CNN
   9 | - Faster R-CNN
  10 | - Mask R-CNN
  11 | - Light-Head R-CNN
  12 | - Cascade R-CNN
  13 | - SPP-Net
  14 | - YOLO
  15 | - YOLOv2
  16 | - YOLOv3
  17 | - YOLT
  18 | - SSD
  19 | - DSSD
  20 | - FSSD
  21 | - ESSD
  22 | - MDSSD
  23 | - Pelee
  24 | - Fire SSD
  25 | - R-FCN
  26 | - FPN
  27 | - DSOD
  28 | - RetinaNet
  29 | - MegDet
  30 | - RefineNet
  31 | - DetNet
  32 | - SSOD
  33 | - CornerNet
  34 | - M2Det
  35 | - 3D Object Detection
  36 | - ZSD（Zero-Shot Object Detection）
  37 | - OSD（One-Shot object Detection）
  38 | - Weakly Supervised Object Detection
  39 | - Softer-NMS
  40 | - 2018
  41 | - 2019
  42 | - Other
  43 | 
  44 | Based on handong1587's github: https://handong1587.github.io/deep_learning/2015/10/09/object-detection.html
  45 | 
  46 | # Survey
  47 | 
  48 | **Imbalance Problems in Object Detection: A Review**
  49 | 
  50 | - intro: under review at TPAMI
  51 | - arXiv: <https://arxiv.org/abs/1909.00169>
  52 | 
  53 | **Recent Advances in Deep Learning for Object Detection**
  54 | 
  55 | - intro: From 2013 (OverFeat) to 2019 (DetNAS)
  56 | - arXiv: <https://arxiv.org/abs/1908.03673>
  57 | 
  58 | **A Survey of Deep Learning-based Object Detection**
  59 | 
  60 | - intro：From Fast R-CNN to NAS-FPN
  61 | 
  62 | - arXiv：<https://arxiv.org/abs/1907.09408>
  63 | 
  64 | **Object Detection in 20 Years: A Survey**
  65 | 
  66 | - intro：This work has been submitted to the IEEE TPAMI for possible publication
  67 | - arXiv：<https://arxiv.org/abs/1905.05055>
  68 | 
  69 | **《Recent Advances in Object Detection in the Age of Deep Convolutional Neural Networks》**
  70 | 
  71 | - intro: awesome
  72 | 
  73 | 
  74 | - arXiv: https://arxiv.org/abs/1809.03193
  75 | 
  76 | **《Deep Learning for Generic Object Detection: A Survey》**
  77 | 
  78 | - intro: Submitted to IJCV 2018
  79 | - arXiv: https://arxiv.org/abs/1809.02165
  80 | 
  81 | # Papers&Codes
  82 | 
  83 | ## R-CNN
  84 | 
  85 | **Rich feature hierarchies for accurate object detection and semantic segmentation**
  86 | 
  87 | - intro: R-CNN
  88 | - arxiv: <http://arxiv.org/abs/1311.2524>
  89 | - supp: <http://people.eecs.berkeley.edu/~rbg/papers/r-cnn-cvpr-supp.pdf>
  90 | - slides: <http://www.image-net.org/challenges/LSVRC/2013/slides/r-cnn-ilsvrc2013-workshop.pdf>
  91 | - slides: <http://www.cs.berkeley.edu/~rbg/slides/rcnn-cvpr14-slides.pdf>
  92 | - github: <https://github.com/rbgirshick/rcnn>
  93 | - notes: <http://zhangliliang.com/2014/07/23/paper-note-rcnn/>
  94 | - caffe-pr("Make R-CNN the Caffe detection example"): <https://github.com/BVLC/caffe/pull/482>
  95 | 
  96 | ## Fast R-CNN
  97 | 
  98 | **Fast R-CNN**
  99 | 
 100 | - arxiv: <http://arxiv.org/abs/1504.08083>
 101 | - slides: <http://tutorial.caffe.berkeleyvision.org/caffe-cvpr15-detection.pdf>
 102 | - github: <https://github.com/rbgirshick/fast-rcnn>
 103 | - github(COCO-branch): <https://github.com/rbgirshick/fast-rcnn/tree/coco>
 104 | - webcam demo: <https://github.com/rbgirshick/fast-rcnn/pull/29>
 105 | - notes: <http://zhangliliang.com/2015/05/17/paper-note-fast-rcnn/>
 106 | - notes: <http://blog.csdn.net/linj_m/article/details/48930179>
 107 | - github("Fast R-CNN in MXNet"): <https://github.com/precedenceguo/mx-rcnn>
 108 | - github: <https://github.com/mahyarnajibi/fast-rcnn-torch>
 109 | - github: <https://github.com/apple2373/chainer-simple-fast-rnn>
 110 | - github: <https://github.com/zplizzi/tensorflow-fast-rcnn>
 111 | 
 112 | **A-Fast-RCNN: Hard Positive Generation via Adversary for Object Detection**
 113 | 
 114 | - intro: CVPR 2017
 115 | - arxiv: <https://arxiv.org/abs/1704.03414>
 116 | - paper: <http://abhinavsh.info/papers/pdfs/adversarial_object_detection.pdf>
 117 | - github(Caffe): <https://github.com/xiaolonw/adversarial-frcnn>
 118 | 
 119 | ## Faster R-CNN
 120 | 
 121 | **Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks**
 122 | 
 123 | - intro: NIPS 2015
 124 | - arxiv: <http://arxiv.org/abs/1506.01497>
 125 | - gitxiv: <http://www.gitxiv.com/posts/8pfpcvefDYn2gSgXk/faster-r-cnn-towards-real-time-object-detection-with-region>
 126 | - slides: <http://web.cs.hacettepe.edu.tr/~aykut/classes/spring2016/bil722/slides/w05-FasterR-CNN.pdf>
 127 | - github(official, Matlab): <https://github.com/ShaoqingRen/faster_rcnn>
 128 | - github(Caffe): <https://github.com/rbgirshick/py-faster-rcnn>
 129 | - github(MXNet): <https://github.com/msracver/Deformable-ConvNets/tree/master/faster_rcnn>
 130 | - github(PyTorch--recommend): <https://github.com//jwyang/faster-rcnn.pytorch>
 131 | - github: <https://github.com/mitmul/chainer-faster-rcnn>
 132 | - github(Torch):: <https://github.com/andreaskoepf/faster-rcnn.torch>
 133 | - github(Torch):: <https://github.com/ruotianluo/Faster-RCNN-Densecap-torch>
 134 | - github(TensorFlow): <https://github.com/smallcorgi/Faster-RCNN_TF>
 135 | - github(TensorFlow): <https://github.com/CharlesShang/TFFRCNN>
 136 | - github(C++ demo): <https://github.com/YihangLou/FasterRCNN-Encapsulation-Cplusplus>
 137 | - github(Keras): <https://github.com/yhenon/keras-frcnn>
 138 | - github: <https://github.com/Eniac-Xie/faster-rcnn-resnet>
 139 | - github(C++): <https://github.com/D-X-Y/caffe-faster-rcnn/tree/dev>
 140 | 
 141 | **R-CNN minus R**
 142 | 
 143 | - intro: BMVC 2015
 144 | - arxiv: <http://arxiv.org/abs/1506.06981>
 145 | 
 146 | **Faster R-CNN in MXNet with distributed implementation and data parallelization**
 147 | 
 148 | - github: <https://github.com/dmlc/mxnet/tree/master/example/rcnn>
 149 | 
 150 | **Contextual Priming and Feedback for Faster R-CNN**
 151 | 
 152 | - intro: ECCV 2016. Carnegie Mellon University
 153 | - paper: <http://abhinavsh.info/context_priming_feedback.pdf>
 154 | - poster: <http://www.eccv2016.org/files/posters/P-1A-20.pdf>
 155 | 
 156 | **An Implementation of Faster RCNN with Study for Region Sampling**
 157 | 
 158 | - intro: Technical Report, 3 pages. CMU
 159 | - arxiv: <https://arxiv.org/abs/1702.02138>
 160 | - github: <https://github.com/endernewton/tf-faster-rcnn>
 161 | - github: https://github.com/ruotianluo/pytorch-faster-rcnn
 162 | 
 163 | **Interpretable R-CNN**
 164 | 
 165 | - intro: North Carolina State University & Alibaba
 166 | - keywords: AND-OR Graph (AOG)
 167 | - arxiv: <https://arxiv.org/abs/1711.05226>
 168 | 
 169 | **Domain Adaptive Faster R-CNN for Object Detection in the Wild**
 170 | 
 171 | - intro: CVPR 2018. ETH Zurich & ESAT/PSI
 172 | - arxiv: <https://arxiv.org/abs/1803.03243>
 173 | 
 174 | ## Mask R-CNN
 175 | 
 176 | - arxiv: <http://arxiv.org/abs/1703.06870>
 177 | - github(Keras): https://github.com/matterport/Mask_RCNN
 178 | - github(Caffe2): https://github.com/facebookresearch/Detectron
 179 | - github(Pytorch): <https://github.com/wannabeOG/Mask-RCNN>
 180 | - github(MXNet): https://github.com/TuSimple/mx-maskrcnn
 181 | - github(Chainer): https://github.com/DeNA/Chainer_Mask_R-CNN
 182 | 
 183 | ## Light-Head R-CNN
 184 | 
 185 | **Light-Head R-CNN: In Defense of Two-Stage Object Detector**
 186 | 
 187 | - intro: Tsinghua University & Megvii Inc
 188 | - arxiv: <https://arxiv.org/abs/1711.07264>
 189 | - github(offical): https://github.com/zengarden/light_head_rcnn
 190 | - github: <https://github.com/terrychenism/Deformable-ConvNets/blob/master/rfcn/symbols/resnet_v1_101_rfcn_light.py#L784>
 191 | 
 192 | ## Cascade R-CNN
 193 | 
 194 | **Cascade R-CNN: Delving into High Quality Object Detection**
 195 | 
 196 | - arxiv: <https://arxiv.org/abs/1712.00726>
 197 | - github: <https://github.com/zhaoweicai/cascade-rcnn>
 198 | 
 199 | ## SPP-Net
 200 | 
 201 | **Spatial Pyramid Pooling in Deep Convolutional Networks for Visual Recognition**
 202 | 
 203 | - intro: ECCV 2014 / TPAMI 2015
 204 | - arxiv: <http://arxiv.org/abs/1406.4729>
 205 | - github: <https://github.com/ShaoqingRen/SPP_net>
 206 | - notes: <http://zhangliliang.com/2014/09/13/paper-note-sppnet/>
 207 | 
 208 | **DeepID-Net: Deformable Deep Convolutional Neural Networks for Object Detection**
 209 | 
 210 | - intro: PAMI 2016
 211 | - intro: an extension of R-CNN. box pre-training, cascade on region proposals, deformation layers and context representations
 212 | - project page: <http://www.ee.cuhk.edu.hk/%CB%9Cwlouyang/projects/imagenetDeepId/index.html>
 213 | - arxiv: <http://arxiv.org/abs/1412.5661>
 214 | 
 215 | **Object Detectors Emerge in Deep Scene CNNs**
 216 | 
 217 | - intro: ICLR 2015
 218 | - arxiv: <http://arxiv.org/abs/1412.6856>
 219 | - paper: <https://www.robots.ox.ac.uk/~vgg/rg/papers/zhou_iclr15.pdf>
 220 | - paper: <https://people.csail.mit.edu/khosla/papers/iclr2015_zhou.pdf>
 221 | - slides: <http://places.csail.mit.edu/slide_iclr2015.pdf>
 222 | 
 223 | **segDeepM: Exploiting Segmentation and Context in Deep Neural Networks for Object Detection**
 224 | 
 225 | - intro: CVPR 2015
 226 | - project(code+data): <https://www.cs.toronto.edu/~yukun/segdeepm.html>
 227 | - arxiv: <https://arxiv.org/abs/1502.04275>
 228 | - github: <https://github.com/YknZhu/segDeepM>
 229 | 
 230 | **Object Detection Networks on Convolutional Feature Maps**
 231 | 
 232 | - intro: TPAMI 2015
 233 | - keywords: NoC
 234 | - arxiv: <http://arxiv.org/abs/1504.06066>
 235 | 
 236 | **Improving Object Detection with Deep Convolutional Networks via Bayesian Optimization and Structured Prediction**
 237 | 
 238 | - arxiv: <http://arxiv.org/abs/1504.03293>
 239 | - slides: <http://www.ytzhang.net/files/publications/2015-cvpr-det-slides.pdf>
 240 | - github: <https://github.com/YutingZhang/fgs-obj>
 241 | 
 242 | **DeepBox: Learning Objectness with Convolutional Networks**
 243 | 
 244 | - keywords: DeepBox
 245 | - arxiv: <http://arxiv.org/abs/1505.02146>
 246 | - github: <https://github.com/weichengkuo/DeepBox>
 247 | 
 248 | ## YOLO
 249 | 
 250 | **You Only Look Once: Unified, Real-Time Object Detection**
 251 | 
 252 | [![img](https://camo.githubusercontent.com/e69d4118b20a42de4e23b9549f9a6ec6dbbb0814/687474703a2f2f706a7265646469652e636f6d2f6d656469612f66696c65732f6461726b6e65742d626c61636b2d736d616c6c2e706e67)](https://camo.githubusercontent.com/e69d4118b20a42de4e23b9549f9a6ec6dbbb0814/687474703a2f2f706a7265646469652e636f6d2f6d656469612f66696c65732f6461726b6e65742d626c61636b2d736d616c6c2e706e67)
 253 | 
 254 | - arxiv: <http://arxiv.org/abs/1506.02640>
 255 | - code: <https://pjreddie.com/darknet/yolov1/>
 256 | - github: <https://github.com/pjreddie/darknet>
 257 | - blog: <https://pjreddie.com/darknet/yolov1/>
 258 | - slides: <https://docs.google.com/presentation/d/1aeRvtKG21KHdD5lg6Hgyhx5rPq_ZOsGjG5rJ1HP7BbA/pub?start=false&loop=false&delayms=3000&slide=id.p>
 259 | - reddit: <https://www.reddit.com/r/MachineLearning/comments/3a3m0o/realtime_object_detection_with_yolo/>
 260 | - github: <https://github.com/gliese581gg/YOLO_tensorflow>
 261 | - github: <https://github.com/xingwangsfu/caffe-yolo>
 262 | - github: <https://github.com/frankzhangrui/Darknet-Yolo>
 263 | - github: <https://github.com/BriSkyHekun/py-darknet-yolo>
 264 | - github: <https://github.com/tommy-qichang/yolo.torch>
 265 | - github: <https://github.com/frischzenger/yolo-windows>
 266 | - github: <https://github.com/AlexeyAB/yolo-windows>
 267 | - github: <https://github.com/nilboy/tensorflow-yolo>
 268 | 
 269 | **darkflow - translate darknet to tensorflow. Load trained weights, retrain/fine-tune them using tensorflow, export constant graph def to C++**
 270 | 
 271 | - blog: <https://thtrieu.github.io/notes/yolo-tensorflow-graph-buffer-cpp>
 272 | - github: <https://github.com/thtrieu/darkflow>
 273 | 
 274 | **Start Training YOLO with Our Own Data**
 275 | 
 276 | [![img](https://camo.githubusercontent.com/2f99b692dd7ce47d7832385f3e8a6654e680d92a/687474703a2f2f6775616e6768616e2e696e666f2f626c6f672f656e2f77702d636f6e74656e742f75706c6f6164732f323031352f31322f696d616765732d34302e6a7067)](https://camo.githubusercontent.com/2f99b692dd7ce47d7832385f3e8a6654e680d92a/687474703a2f2f6775616e6768616e2e696e666f2f626c6f672f656e2f77702d636f6e74656e742f75706c6f6164732f323031352f31322f696d616765732d34302e6a7067)
 277 | 
 278 | - intro: train with customized data and class numbers/labels. Linux / Windows version for darknet.
 279 | - blog: <http://guanghan.info/blog/en/my-works/train-yolo/>
 280 | - github: <https://github.com/Guanghan/darknet>
 281 | 
 282 | **YOLO: Core ML versus MPSNNGraph**
 283 | 
 284 | - intro: Tiny YOLO for iOS implemented using CoreML but also using the new MPS graph API.
 285 | - blog: <http://machinethink.net/blog/yolo-coreml-versus-mps-graph/>
 286 | - github: <https://github.com/hollance/YOLO-CoreML-MPSNNGraph>
 287 | 
 288 | **TensorFlow YOLO object detection on Android**
 289 | 
 290 | - intro: Real-time object detection on Android using the YOLO network with TensorFlow
 291 | - github: <https://github.com/natanielruiz/android-yolo>
 292 | 
 293 | **Computer Vision in iOS – Object Detection**
 294 | 
 295 | - blog: <https://sriraghu.com/2017/07/12/computer-vision-in-ios-object-detection/>
 296 | - github:<https://github.com/r4ghu/iOS-CoreML-Yolo>
 297 | 
 298 | ## YOLOv2
 299 | 
 300 | **YOLO9000: Better, Faster, Stronger**
 301 | 
 302 | - arxiv: <https://arxiv.org/abs/1612.08242>
 303 | - code: <http://pjreddie.com/yolo9000/>    https://pjreddie.com/darknet/yolov2/
 304 | - github(Chainer): <https://github.com/leetenki/YOLOv2>
 305 | - github(Keras): <https://github.com/allanzelener/YAD2K>
 306 | - github(PyTorch): <https://github.com/longcw/yolo2-pytorch>
 307 | - github(Tensorflow): <https://github.com/hizhangp/yolo_tensorflow>
 308 | - github(Windows): <https://github.com/AlexeyAB/darknet>
 309 | - github: <https://github.com/choasUp/caffe-yolo9000>
 310 | - github: <https://github.com/philipperemy/yolo-9000>
 311 | - github(TensorFlow): <https://github.com/KOD-Chen/YOLOv2-Tensorflow>
 312 | - github(Keras): <https://github.com/yhcc/yolo2>
 313 | - github(Keras): <https://github.com/experiencor/keras-yolo2>
 314 | - github(TensorFlow): <https://github.com/WojciechMormul/yolo2>
 315 | 
 316 | **darknet_scripts**
 317 | 
 318 | - intro: Auxilary scripts to work with (YOLO) darknet deep learning famework. AKA -> How to generate YOLO anchors?
 319 | - github: <https://github.com/Jumabek/darknet_scripts>
 320 | 
 321 | **Yolo_mark: GUI for marking bounded boxes of objects in images for training Yolo v2**
 322 | 
 323 | - github: <https://github.com/AlexeyAB/Yolo_mark>
 324 | 
 325 | **LightNet: Bringing pjreddie's DarkNet out of the shadows**
 326 | 
 327 | <https://github.com//explosion/lightnet>
 328 | 
 329 | **YOLO v2 Bounding Box Tool**
 330 | 
 331 | - intro: Bounding box labeler tool to generate the training data in the format YOLO v2 requires.
 332 | - github: <https://github.com/Cartucho/yolo-boundingbox-labeler-GUI>
 333 | 
 334 | **Loss Rank Mining: A General Hard Example Mining Method for Real-time Detectors**
 335 | 
 336 | - intro: **LRM** is the first hard example mining strategy which could fit YOLOv2 perfectly and make it better applied in series of real scenarios where both real-time rates and accurate detection are strongly demanded.
 337 | - arxiv: https://arxiv.org/abs/1804.04606
 338 | 
 339 | **Object detection at 200 Frames Per Second**
 340 | 
 341 | - intro: faster than Tiny-Yolo-v2
 342 | - arxiv: https://arxiv.org/abs/1805.06361
 343 | 
 344 | **Event-based Convolutional Networks for Object Detection in Neuromorphic Cameras**
 345 | 
 346 | - intro: YOLE--Object Detection in Neuromorphic Cameras
 347 | - arxiv:https://arxiv.org/abs/1805.07931
 348 | 
 349 | **OmniDetector: With Neural Networks to Bounding Boxes**
 350 | 
 351 | - intro: a person detector on n fish-eye images of indoor scenes（NIPS 2018）
 352 | - arxiv:https://arxiv.org/abs/1805.08503
 353 | - datasets:https://gitlab.com/omnidetector/omnidetector
 354 | 
 355 | ## YOLOv3
 356 | 
 357 | **YOLOv3: An Incremental Improvement**
 358 | 
 359 | - arxiv:https://arxiv.org/abs/1804.02767
 360 | - paper:https://pjreddie.com/media/files/papers/YOLOv3.pdf
 361 | - code: <https://pjreddie.com/darknet/yolo/>
 362 | - github(Official):https://github.com/pjreddie/darknet
 363 | - github:https://github.com/mystic123/tensorflow-yolo-v3
 364 | - github:https://github.com/experiencor/keras-yolo3
 365 | - github:https://github.com/qqwweee/keras-yolo3
 366 | - github:https://github.com/marvis/pytorch-yolo3
 367 | - github:https://github.com/ayooshkathuria/pytorch-yolo-v3
 368 | - github:https://github.com/ayooshkathuria/YOLO_v3_tutorial_from_scratch
 369 | - github:https://github.com/eriklindernoren/PyTorch-YOLOv3
 370 | - github:https://github.com/ultralytics/yolov3
 371 | - github:https://github.com/BobLiu20/YOLOv3_PyTorch
 372 | - github:https://github.com/andy-yun/pytorch-0.4-yolov3
 373 | - github:https://github.com/DeNA/PyTorch_YOLOv3
 374 | 
 375 | ## YOLT
 376 | 
 377 | **You Only Look Twice: Rapid Multi-Scale Object Detection In Satellite Imagery**
 378 | 
 379 | - intro: Small Object Detection
 380 | 
 381 | 
 382 | - arxiv:https://arxiv.org/abs/1805.09512
 383 | - github:https://github.com/avanetten/yolt
 384 | 
 385 | ## SSD
 386 | 
 387 | **SSD: Single Shot MultiBox Detector**
 388 | 
 389 | [![img](https://camo.githubusercontent.com/ad9b147ed3a5f48ffb7c3540711c15aa04ce49c6/687474703a2f2f7777772e63732e756e632e6564752f7e776c69752f7061706572732f7373642e706e67)](https://camo.githubusercontent.com/ad9b147ed3a5f48ffb7c3540711c15aa04ce49c6/687474703a2f2f7777772e63732e756e632e6564752f7e776c69752f7061706572732f7373642e706e67)
 390 | 
 391 | - intro: ECCV 2016 Oral
 392 | - arxiv: <http://arxiv.org/abs/1512.02325>
 393 | - paper: <http://www.cs.unc.edu/~wliu/papers/ssd.pdf>
 394 | - slides: [http://www.cs.unc.edu/%7Ewliu/papers/ssd_eccv2016_slide.pdf](http://www.cs.unc.edu/~wliu/papers/ssd_eccv2016_slide.pdf)
 395 | - github(Official): <https://github.com/weiliu89/caffe/tree/ssd>
 396 | - video: <http://weibo.com/p/2304447a2326da963254c963c97fb05dd3a973>
 397 | - github: <https://github.com/zhreshold/mxnet-ssd>
 398 | - github: <https://github.com/zhreshold/mxnet-ssd.cpp>
 399 | - github: <https://github.com/rykov8/ssd_keras>
 400 | - github: <https://github.com/balancap/SSD-Tensorflow>
 401 | - github: <https://github.com/amdegroot/ssd.pytorch>
 402 | - github(Caffe): <https://github.com/chuanqi305/MobileNet-SSD>
 403 | 
 404 | **What's the diffience in performance between this new code you pushed and the previous code? #327**
 405 | 
 406 | <https://github.com/weiliu89/caffe/issues/327>
 407 | 
 408 | ## DSSD
 409 | 
 410 | **DSSD : Deconvolutional Single Shot Detector**
 411 | 
 412 | - intro: UNC Chapel Hill & Amazon Inc
 413 | - arxiv: <https://arxiv.org/abs/1701.06659>
 414 | - github: <https://github.com/chengyangfu/caffe/tree/dssd>
 415 | - github: <https://github.com/MTCloudVision/mxnet-dssd>
 416 | - demo: <http://120.52.72.53/www.cs.unc.edu/c3pr90ntc0td/~cyfu/dssd_lalaland.mp4>
 417 | 
 418 | **Enhancement of SSD by concatenating feature maps for object detection**
 419 | 
 420 | - intro: rainbow SSD (R-SSD)
 421 | - arxiv: <https://arxiv.org/abs/1705.09587>
 422 | 
 423 | **Context-aware Single-Shot Detector**
 424 | 
 425 | - keywords: CSSD, DiCSSD, DeCSSD, effective receptive fields (ERFs), theoretical receptive fields (TRFs)
 426 | - arxiv: <https://arxiv.org/abs/1707.08682>
 427 | 
 428 | **Feature-Fused SSD: Fast Detection for Small Objects**
 429 | 
 430 | <https://arxiv.org/abs/1709.05054>
 431 | 
 432 | ## FSSD
 433 | 
 434 | **FSSD: Feature Fusion Single Shot Multibox Detector**
 435 | 
 436 | <https://arxiv.org/abs/1712.00960>
 437 | 
 438 | **Weaving Multi-scale Context for Single Shot Detector**
 439 | 
 440 | - intro: WeaveNet
 441 | - keywords: fuse multi-scale information
 442 | - arxiv: <https://arxiv.org/abs/1712.03149>
 443 | 
 444 | ## ESSD
 445 | 
 446 | **Extend the shallow part of Single Shot MultiBox Detector via Convolutional Neural Network**
 447 | 
 448 | <https://arxiv.org/abs/1801.05918>
 449 | 
 450 | **Tiny SSD: A Tiny Single-shot Detection Deep Convolutional Neural Network for Real-time Embedded Object Detection**
 451 | 
 452 | <https://arxiv.org/abs/1802.06488>
 453 | 
 454 | ## MDSSD
 455 | 
 456 | **MDSSD: Multi-scale Deconvolutional Single Shot Detector for small objects**
 457 | 
 458 | - arxiv: https://arxiv.org/abs/1805.07009
 459 | 
 460 | ## Pelee
 461 | 
 462 | **Pelee: A Real-Time Object Detection System on Mobile Devices**
 463 | 
 464 | https://github.com/Robert-JunWang/Pelee
 465 | 
 466 | - intro: (ICLR 2018 workshop track)
 467 | 
 468 | 
 469 | - arxiv: https://arxiv.org/abs/1804.06882
 470 | - github: https://github.com/Robert-JunWang/Pelee
 471 | 
 472 | ## Fire SSD
 473 | 
 474 | **Fire SSD: Wide Fire Modules based Single Shot Detector on Edge Device**
 475 | 
 476 | - intro:low cost, fast speed and high mAP on  factor edge computing devices
 477 | 
 478 | 
 479 | - arxiv:https://arxiv.org/abs/1806.05363
 480 | 
 481 | ## R-FCN
 482 | 
 483 | **R-FCN: Object Detection via Region-based Fully Convolutional Networks**
 484 | 
 485 | - arxiv: <http://arxiv.org/abs/1605.06409>
 486 | - github: <https://github.com/daijifeng001/R-FCN>
 487 | - github(MXNet): <https://github.com/msracver/Deformable-ConvNets/tree/master/rfcn>
 488 | - github: <https://github.com/Orpine/py-R-FCN>
 489 | - github: <https://github.com/PureDiors/pytorch_RFCN>
 490 | - github: <https://github.com/bharatsingh430/py-R-FCN-multiGPU>
 491 | - github: <https://github.com/xdever/RFCN-tensorflow>
 492 | 
 493 | **R-FCN-3000 at 30fps: Decoupling Detection and Classification**
 494 | 
 495 | <https://arxiv.org/abs/1712.01802>
 496 | 
 497 | **Recycle deep features for better object detection**
 498 | 
 499 | - arxiv: <http://arxiv.org/abs/1607.05066>
 500 | 
 501 | ## FPN
 502 | 
 503 | **Feature Pyramid Networks for Object Detection**
 504 | 
 505 | - intro: Facebook AI Research
 506 | - arxiv: <https://arxiv.org/abs/1612.03144>
 507 | 
 508 | **Action-Driven Object Detection with Top-Down Visual Attentions**
 509 | 
 510 | - arxiv: <https://arxiv.org/abs/1612.06704>
 511 | 
 512 | **Beyond Skip Connections: Top-Down Modulation for Object Detection**
 513 | 
 514 | - intro: CMU & UC Berkeley & Google Research
 515 | - arxiv: <https://arxiv.org/abs/1612.06851>
 516 | 
 517 | **Wide-Residual-Inception Networks for Real-time Object Detection**
 518 | 
 519 | - intro: Inha University
 520 | - arxiv: <https://arxiv.org/abs/1702.01243>
 521 | 
 522 | **Attentional Network for Visual Object Detection**
 523 | 
 524 | - intro: University of Maryland & Mitsubishi Electric Research Laboratories
 525 | - arxiv: <https://arxiv.org/abs/1702.01478>
 526 | 
 527 | **Learning Chained Deep Features and Classifiers for Cascade in Object Detection**
 528 | 
 529 | - keykwords: CC-Net
 530 | - intro: chained cascade network (CC-Net). 81.1% mAP on PASCAL VOC 2007
 531 | - arxiv: <https://arxiv.org/abs/1702.07054>
 532 | 
 533 | **DeNet: Scalable Real-time Object Detection with Directed Sparse Sampling**
 534 | 
 535 | - intro: ICCV 2017 (poster)
 536 | - arxiv: <https://arxiv.org/abs/1703.10295>
 537 | 
 538 | **Discriminative Bimodal Networks for Visual Localization and Detection with Natural Language Queries**
 539 | 
 540 | - intro: CVPR 2017
 541 | - arxiv: <https://arxiv.org/abs/1704.03944>
 542 | 
 543 | **Spatial Memory for Context Reasoning in Object Detection**
 544 | 
 545 | - arxiv: <https://arxiv.org/abs/1704.04224>
 546 | 
 547 | **Accurate Single Stage Detector Using Recurrent Rolling Convolution**
 548 | 
 549 | - intro: CVPR 2017. SenseTime
 550 | - keywords: Recurrent Rolling Convolution (RRC)
 551 | - arxiv: <https://arxiv.org/abs/1704.05776>
 552 | - github: <https://github.com/xiaohaoChen/rrc_detection>
 553 | 
 554 | **Deep Occlusion Reasoning for Multi-Camera Multi-Target Detection**
 555 | 
 556 | <https://arxiv.org/abs/1704.05775>
 557 | 
 558 | **LCDet: Low-Complexity Fully-Convolutional Neural Networks for Object Detection in Embedded Systems**
 559 | 
 560 | - intro: Embedded Vision Workshop in CVPR. UC San Diego & Qualcomm Inc
 561 | - arxiv: <https://arxiv.org/abs/1705.05922>
 562 | 
 563 | **Point Linking Network for Object Detection**
 564 | 
 565 | - intro: Point Linking Network (PLN)
 566 | - arxiv: <https://arxiv.org/abs/1706.03646>
 567 | 
 568 | **Perceptual Generative Adversarial Networks for Small Object Detection**
 569 | 
 570 | <https://arxiv.org/abs/1706.05274>
 571 | 
 572 | **Few-shot Object Detection**
 573 | 
 574 | <https://arxiv.org/abs/1706.08249>
 575 | 
 576 | **Yes-Net: An effective Detector Based on Global Information**
 577 | 
 578 | <https://arxiv.org/abs/1706.09180>
 579 | 
 580 | **SMC Faster R-CNN: Toward a scene-specialized multi-object detector**
 581 | 
 582 | <https://arxiv.org/abs/1706.10217>
 583 | 
 584 | **Towards lightweight convolutional neural networks for object detection**
 585 | 
 586 | <https://arxiv.org/abs/1707.01395>
 587 | 
 588 | **RON: Reverse Connection with Objectness Prior Networks for Object Detection**
 589 | 
 590 | - intro: CVPR 2017
 591 | - arxiv: <https://arxiv.org/abs/1707.01691>
 592 | - github: <https://github.com/taokong/RON>
 593 | 
 594 | **Mimicking Very Efficient Network for Object Detection**
 595 | 
 596 | - intro: CVPR 2017. SenseTime & Beihang University
 597 | - paper: <http://openaccess.thecvf.com/content_cvpr_2017/papers/Li_Mimicking_Very_Efficient_CVPR_2017_paper.pdf>
 598 | 
 599 | **Residual Features and Unified Prediction Network for Single Stage Detection**
 600 | 
 601 | <https://arxiv.org/abs/1707.05031>
 602 | 
 603 | **Deformable Part-based Fully Convolutional Network for Object Detection**
 604 | 
 605 | - intro: BMVC 2017 (oral). Sorbonne Universités & CEDRIC
 606 | - arxiv: <https://arxiv.org/abs/1707.06175>
 607 | 
 608 | **Adaptive Feeding: Achieving Fast and Accurate Detections by Adaptively Combining Object Detectors**
 609 | 
 610 | - intro: ICCV 2017
 611 | - arxiv: <https://arxiv.org/abs/1707.06399>
 612 | 
 613 | **Recurrent Scale Approximation for Object Detection in CNN**
 614 | 
 615 | - intro: ICCV 2017
 616 | - keywords: Recurrent Scale Approximation (RSA)
 617 | - arxiv: <https://arxiv.org/abs/1707.09531>
 618 | - github: <https://github.com/sciencefans/RSA-for-object-detection>
 619 | 
 620 | ## DSOD
 621 | 
 622 | **DSOD: Learning Deeply Supervised Object Detectors from Scratch**
 623 | 
 624 | ![img](https://user-images.githubusercontent.com/3794909/28934967-718c9302-78b5-11e7-89ee-8b514e53e23c.png)
 625 | 
 626 | - intro: ICCV 2017. Fudan University & Tsinghua University & Intel Labs China
 627 | - arxiv: <https://arxiv.org/abs/1708.01241>
 628 | - github: <https://github.com/szq0214/DSOD>
 629 | - github:https://github.com/Windaway/DSOD-Tensorflow
 630 | - github:https://github.com/chenyuntc/dsod.pytorch
 631 | 
 632 | **Learning Object Detectors from Scratch with Gated Recurrent Feature Pyramids**
 633 | 
 634 | - arxiv:https://arxiv.org/abs/1712.00886
 635 | - github:https://github.com/szq0214/GRP-DSOD
 636 | 
 637 | **Tiny-DSOD: Lightweight Object Detection for Resource-Restricted Usages**
 638 | 
 639 | - intro: BMVC 2018
 640 | - arXiv: https://arxiv.org/abs/1807.11013
 641 | 
 642 | **Object Detection from Scratch with Deep Supervision**
 643 | 
 644 | - intro: This is an extended version of DSOD
 645 | - arXiv: https://arxiv.org/abs/1809.09294
 646 | 
 647 | ## RetinaNet
 648 | 
 649 | **Focal Loss for Dense Object Detection**
 650 | 
 651 | - intro: ICCV 2017 Best student paper award. Facebook AI Research
 652 | - keywords: RetinaNet
 653 | - arxiv: <https://arxiv.org/abs/1708.02002>
 654 | 
 655 | **CoupleNet: Coupling Global Structure with Local Parts for Object Detection**
 656 | 
 657 | - intro: ICCV 2017
 658 | - arxiv: <https://arxiv.org/abs/1708.02863>
 659 | 
 660 | **Incremental Learning of Object Detectors without Catastrophic Forgetting**
 661 | 
 662 | - intro: ICCV 2017. Inria
 663 | - arxiv: <https://arxiv.org/abs/1708.06977>
 664 | 
 665 | **Zoom Out-and-In Network with Map Attention Decision for Region Proposal and Object Detection**
 666 | 
 667 | <https://arxiv.org/abs/1709.04347>
 668 | 
 669 | **StairNet: Top-Down Semantic Aggregation for Accurate One Shot Detection**
 670 | 
 671 | <https://arxiv.org/abs/1709.05788>
 672 | 
 673 | **Dynamic Zoom-in Network for Fast Object Detection in Large Images**
 674 | 
 675 | <https://arxiv.org/abs/1711.05187>
 676 | 
 677 | **Zero-Annotation Object Detection with Web Knowledge Transfer**
 678 | 
 679 | - intro: NTU, Singapore & Amazon
 680 | - keywords: multi-instance multi-label domain adaption learning framework
 681 | - arxiv: <https://arxiv.org/abs/1711.05954>
 682 | 
 683 | ## MegDet
 684 | 
 685 | **MegDet: A Large Mini-Batch Object Detector**
 686 | 
 687 | - intro: Peking University & Tsinghua University & Megvii Inc
 688 | - arxiv: <https://arxiv.org/abs/1711.07240>
 689 | 
 690 | **Receptive Field Block Net for Accurate and Fast Object Detection**
 691 | 
 692 | - intro: RFBNet
 693 | - arxiv: <https://arxiv.org/abs/1711.07767>
 694 | - github: <https://github.com//ruinmessi/RFBNet>
 695 | 
 696 | **An Analysis of Scale Invariance in Object Detection - SNIP**
 697 | 
 698 | - arxiv: <https://arxiv.org/abs/1711.08189>
 699 | - github: <https://github.com/bharatsingh430/snip>
 700 | 
 701 | **Feature Selective Networks for Object Detection**
 702 | 
 703 | <https://arxiv.org/abs/1711.08879>
 704 | 
 705 | **Learning a Rotation Invariant Detector with Rotatable Bounding Box**
 706 | 
 707 | - arxiv: <https://arxiv.org/abs/1711.09405>
 708 | - github: <https://github.com/liulei01/DRBox>
 709 | 
 710 | **Scalable Object Detection for Stylized Objects**
 711 | 
 712 | - intro: Microsoft AI & Research Munich
 713 | - arxiv: <https://arxiv.org/abs/1711.09822>
 714 | 
 715 | **Learning Object Detectors from Scratch with Gated Recurrent Feature Pyramids**
 716 | 
 717 | - arxiv: <https://arxiv.org/abs/1712.00886>
 718 | - github: <https://github.com/szq0214/GRP-DSOD>
 719 | 
 720 | **Deep Regionlets for Object Detection**
 721 | 
 722 | - keywords: region selection network, gating network
 723 | - arxiv: <https://arxiv.org/abs/1712.02408>
 724 | 
 725 | **Training and Testing Object Detectors with Virtual Images**
 726 | 
 727 | - intro: IEEE/CAA Journal of Automatica Sinica
 728 | - arxiv: <https://arxiv.org/abs/1712.08470>
 729 | 
 730 | **Large-Scale Object Discovery and Detector Adaptation from Unlabeled Video**
 731 | 
 732 | - keywords: object mining, object tracking, unsupervised object discovery by appearance-based clustering, self-supervised detector adaptation
 733 | - arxiv: <https://arxiv.org/abs/1712.08832>
 734 | 
 735 | **Spot the Difference by Object Detection**
 736 | 
 737 | - intro: Tsinghua University & JD Group
 738 | - arxiv: <https://arxiv.org/abs/1801.01051>
 739 | 
 740 | **Localization-Aware Active Learning for Object Detection**
 741 | 
 742 | - arxiv: <https://arxiv.org/abs/1801.05124>
 743 | 
 744 | **Object Detection with Mask-based Feature Encoding**
 745 | 
 746 | - arxiv: <https://arxiv.org/abs/1802.03934>
 747 | 
 748 | **LSTD: A Low-Shot Transfer Detector for Object Detection**
 749 | 
 750 | - intro: AAAI 2018
 751 | - arxiv: <https://arxiv.org/abs/1803.01529>
 752 | 
 753 | **Pseudo Mask Augmented Object Detection**
 754 | 
 755 | <https://arxiv.org/abs/1803.05858>
 756 | 
 757 | **Revisiting RCNN: On Awakening the Classification Power of Faster RCNN**
 758 | 
 759 | <https://arxiv.org/abs/1803.06799>
 760 | 
 761 | **Learning Region Features for Object Detection**
 762 | 
 763 | - intro: Peking University & MSRA
 764 | - arxiv: <https://arxiv.org/abs/1803.07066>
 765 | 
 766 | **Single-Shot Bidirectional Pyramid Networks for High-Quality Object Detection**
 767 | 
 768 | - intro: Singapore Management University & Zhejiang University
 769 | - arxiv: <https://arxiv.org/abs/1803.08208>
 770 | 
 771 | **Object Detection for Comics using Manga109 Annotations**
 772 | 
 773 | - intro: University of Tokyo & National Institute of Informatics, Japan
 774 | - arxiv: <https://arxiv.org/abs/1803.08670>
 775 | 
 776 | **Task-Driven Super Resolution: Object Detection in Low-resolution Images**
 777 | 
 778 | - arxiv: <https://arxiv.org/abs/1803.11316>
 779 | 
 780 | **Transferring Common-Sense Knowledge for Object Detection**
 781 | 
 782 | - arxiv: <https://arxiv.org/abs/1804.01077>
 783 | 
 784 | **Multi-scale Location-aware Kernel Representation for Object Detection**
 785 | 
 786 | - intro: CVPR 2018
 787 | - arxiv: <https://arxiv.org/abs/1804.00428>
 788 | - github: <https://github.com/Hwang64/MLKP>
 789 | 
 790 | 
 791 | **Loss Rank Mining: A General Hard Example Mining Method for Real-time Detectors**
 792 | 
 793 | - intro: National University of Defense Technology
 794 | - arxiv: https://arxiv.org/abs/1804.04606
 795 | 
 796 | **Robust Physical Adversarial Attack on Faster R-CNN Object Detector**
 797 | 
 798 | - arxiv: https://arxiv.org/abs/1804.05810
 799 | 
 800 | ## RefineNet
 801 | 
 802 | **Single-Shot Refinement Neural Network for Object Detection**
 803 | 
 804 | - intro: CVPR 2018
 805 | 
 806 | - arxiv: <https://arxiv.org/abs/1711.06897>
 807 | - github: <https://github.com/sfzhang15/RefineDet>
 808 | - github: https://github.com/lzx1413/PytorchSSD
 809 | - github: https://github.com/ddlee96/RefineDet_mxnet
 810 | - github: https://github.com/MTCloudVision/RefineDet-Mxnet
 811 | 
 812 | ## DetNet
 813 | 
 814 | **DetNet: A Backbone network for Object Detection**
 815 | 
 816 | - intro: Tsinghua University & Face++
 817 | - arxiv: https://arxiv.org/abs/1804.06215
 818 | 
 819 | 
 820 | ## SSOD
 821 | 
 822 | **Self-supervisory Signals for Object Discovery and Detection**
 823 | 
 824 | - Google Brain
 825 | - arxiv:https://arxiv.org/abs/1806.03370
 826 | 
 827 | ## CornerNet
 828 | 
 829 | **CornerNet: Detecting Objects as Paired Keypoints**
 830 | 
 831 | - intro: ECCV 2018
 832 | - arXiv: https://arxiv.org/abs/1808.01244
 833 | - github: <https://github.com/umich-vl/CornerNet>
 834 | 
 835 | ## M2Det
 836 | 
 837 | **M2Det: A Single-Shot Object Detector based on Multi-Level Feature Pyramid Network**
 838 | 
 839 | - intro: AAAI 2019
 840 | - arXiv: https://arxiv.org/abs/1811.04533
 841 | - github: https://github.com/qijiezhao/M2Det
 842 | 
 843 | ## 3D Object Detection
 844 | 
 845 | **3D Backbone Network for 3D Object Detection**
 846 | 
 847 | - arXiv: https://arxiv.org/abs/1901.08373
 848 | 
 849 | **LMNet: Real-time Multiclass Object Detection on CPU using 3D LiDARs**
 850 | 
 851 | - arxiv: https://arxiv.org/abs/1805.04902
 852 | - github: https://github.com/CPFL/Autoware/tree/feature/cnn_lidar_detection
 853 | 
 854 | 
 855 | ## ZSD（Zero-Shot Object Detection）
 856 | 
 857 | **Zero-Shot Detection**
 858 | 
 859 | - intro: Australian National University
 860 | - keywords: YOLO
 861 | - arxiv: <https://arxiv.org/abs/1803.07113>
 862 | 
 863 | **Zero-Shot Object Detection**
 864 | 
 865 | - arxiv: https://arxiv.org/abs/1804.04340
 866 | 
 867 | **Zero-Shot Object Detection: Learning to Simultaneously Recognize and Localize Novel Concepts**
 868 | 
 869 | - arxiv: https://arxiv.org/abs/1803.06049
 870 | 
 871 | **Zero-Shot Object Detection by Hybrid Region Embedding**
 872 | 
 873 | - arxiv: https://arxiv.org/abs/1805.06157
 874 | 
 875 | ## OSD（One-Shot Object Detection）
 876 | 
 877 | **Comparison Network for One-Shot Conditional Object Detection**
 878 | 
 879 | - arXiv: https://arxiv.org/abs/1904.02317
 880 | 
 881 | **One-Shot Object Detection**
 882 | 
 883 | RepMet: Representative-based metric learning for classification and one-shot object detection
 884 | 
 885 | - intro: IBM Research AI
 886 | - arxiv:https://arxiv.org/abs/1806.04728
 887 | - github: TODO
 888 | 
 889 | ## Weakly Supervised Object Detection
 890 | 
 891 | **Weakly Supervised Object Detection in Artworks**
 892 | 
 893 | - intro: ECCV 2018 Workshop Computer Vision for Art Analysis
 894 | - arXiv: https://arxiv.org/abs/1810.02569
 895 | - Datasets: https://wsoda.telecom-paristech.fr/downloads/dataset/IconArt_v1.zip
 896 | 
 897 | **Cross-Domain Weakly-Supervised Object Detection through Progressive Domain Adaptation**
 898 | 
 899 | - intro: CVPR 2018
 900 | - arXiv: https://arxiv.org/abs/1803.11365
 901 | - homepage: https://naoto0804.github.io/cross_domain_detection/
 902 | - paper: http://openaccess.thecvf.com/content_cvpr_2018/html/Inoue_Cross-Domain_Weakly-Supervised_Object_CVPR_2018_paper.html
 903 | - github: https://github.com/naoto0804/cross-domain-detection
 904 | 
 905 | ## Softer-NMS
 906 | 
 907 | **《Softer-NMS: Rethinking Bounding Box Regression for Accurate Object Detection》**
 908 | 
 909 | - intro: CMU & Face++
 910 | - arXiv: https://arxiv.org/abs/1809.08545
 911 | - github: https://github.com/yihui-he/softer-NMS
 912 | 
 913 | ## 2019
 914 | 
 915 | **Feature Selective Anchor-Free Module for Single-Shot Object Detection**
 916 | 
 917 | - intro: CVPR 2019
 918 | 
 919 | - arXiv: https://arxiv.org/abs/1903.00621
 920 | 
 921 | **Object Detection based on Region Decomposition and Assembly**
 922 | 
 923 | - intro: AAAI 2019
 924 | 
 925 | - arXiv: https://arxiv.org/abs/1901.08225
 926 | 
 927 | **Bottom-up Object Detection by Grouping Extreme and Center Points**
 928 | 
 929 | - intro: one stage 43.2% on COCO test-dev
 930 | - arXiv: https://arxiv.org/abs/1901.08043
 931 | - github: https://github.com/xingyizhou/ExtremeNet
 932 | 
 933 | **ORSIm Detector: A Novel Object Detection Framework in Optical Remote Sensing Imagery Using Spatial-Frequency Channel Features**
 934 | 
 935 | - intro: IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
 936 | 
 937 | - arXiv: https://arxiv.org/abs/1901.07925
 938 | 
 939 | **Consistent Optimization for Single-Shot Object Detection**
 940 | 
 941 | - intro: improves RetinaNet from 39.1 AP to 40.1 AP on COCO datase
 942 | 
 943 | - arXiv: https://arxiv.org/abs/1901.06563
 944 | 
 945 | **Learning Pairwise Relationship for Multi-object Detection in Crowded Scenes**
 946 | 
 947 | - arXiv: https://arxiv.org/abs/1901.03796
 948 | 
 949 | **RetinaMask: Learning to predict masks improves state-of-the-art single-shot detection for free**
 950 | 
 951 | - arXiv: https://arxiv.org/abs/1901.03353
 952 | - github: https://github.com/chengyangfu/retinamask
 953 | 
 954 | **Region Proposal by Guided Anchoring**
 955 | 
 956 | - intro: CUHK - SenseTime Joint Lab
 957 | - arXiv: https://arxiv.org/abs/1901.03278
 958 | 
 959 | **Scale-Aware Trident Networks for Object Detection**
 960 | 
 961 | - intro: mAP of **48.4** on the COCO dataset
 962 | - arXiv: https://arxiv.org/abs/1901.01892
 963 | 
 964 | ## 2018
 965 | 
 966 | **Large-Scale Object Detection of Images from Network Cameras in Variable Ambient Lighting Conditions**
 967 | 
 968 | - arXiv: https://arxiv.org/abs/1812.11901
 969 | 
 970 | **Strong-Weak Distribution Alignment for Adaptive Object Detection**
 971 | 
 972 | - arXiv: https://arxiv.org/abs/1812.04798
 973 | 
 974 | **AutoFocus: Efficient Multi-Scale Inference**
 975 | 
 976 | - intro: AutoFocus obtains an **mAP of 47.9%** (68.3% at 50% overlap) on the **COCO test-dev** set while processing **6.4 images per second on a Titan X (Pascal) GPU** 
 977 | - arXiv: https://arxiv.org/abs/1812.01600
 978 | 
 979 | **NOTE-RCNN: NOise Tolerant Ensemble RCNN for Semi-Supervised Object Detection**
 980 | 
 981 | - intro: Google Could
 982 | - arXiv: https://arxiv.org/abs/1812.00124
 983 | 
 984 | **SPLAT: Semantic Pixel-Level Adaptation Transforms for Detection**
 985 | 
 986 | - intro: UC Berkeley
 987 | - arXiv: https://arxiv.org/abs/1812.00929
 988 | 
 989 | **Grid R-CNN**
 990 | 
 991 | - intro: SenseTime
 992 | - arXiv: https://arxiv.org/abs/1811.12030
 993 | 
 994 | **Deformable ConvNets v2: More Deformable, Better Results**
 995 | 
 996 | - intro: Microsoft Research Asia
 997 | 
 998 | - arXiv: https://arxiv.org/abs/1811.11168
 999 | 
1000 | **Anchor Box Optimization for Object Detection**
1001 | 
1002 | - intro: Microsoft Research
1003 | - arXiv: https://arxiv.org/abs/1812.00469
1004 | 
1005 | **Efficient Coarse-to-Fine Non-Local Module for the Detection of Small Objects**
1006 | 
1007 | - intro: https://arxiv.org/abs/1811.12152
1008 | 
1009 | **NOTE-RCNN: NOise Tolerant Ensemble RCNN for Semi-Supervised Object Detection**
1010 | 
1011 | - arXiv: https://arxiv.org/abs/1812.00124
1012 | 
1013 | **Learning RoI Transformer for Detecting Oriented Objects in Aerial Images**
1014 | 
1015 | - arXiv: https://arxiv.org/abs/1812.00155
1016 | 
1017 | **Integrated Object Detection and Tracking with Tracklet-Conditioned Detection**
1018 | 
1019 | - intro: Microsoft Research Asia
1020 | - arXiv: https://arxiv.org/abs/1811.11167
1021 | 
1022 | **Deep Regionlets: Blended Representation and Deep Learning for Generic Object Detection**
1023 | 
1024 | - arXiv: https://arxiv.org/abs/1811.11318
1025 | 
1026 |  **Gradient Harmonized Single-stage Detector**
1027 | 
1028 | - intro: AAAI 2019
1029 | - arXiv: https://arxiv.org/abs/1811.05181
1030 | 
1031 | **CFENet: Object Detection with Comprehensive Feature Enhancement Module**
1032 | 
1033 | - intro: ACCV 2018
1034 | - github: https://github.com/qijiezhao/CFENet
1035 | 
1036 | **DeRPN: Taking a further step toward more general object detection**
1037 | 
1038 | - intro: AAAI 2019
1039 | - arXiv: https://arxiv.org/abs/1811.06700
1040 | - github: https://github.com/HCIILAB/DeRPN
1041 | 
1042 | **Hybrid Knowledge Routed Modules for Large-scale Object Detection**
1043 | 
1044 | - intro: Sun Yat-Sen University & Huawei Noah’s Ark Lab
1045 | - arXiv: https://arxiv.org/abs/1810.12681
1046 | - github: https://github.com/chanyn/HKRM
1047 | 
1048 | **《Receptive Field Block Net for Accurate and Fast Object Detection》**
1049 | 
1050 | - intro: ECCV 2018
1051 | - arXiv: [https://arxiv.org/abs/1711.07767](https://arxiv.org/abs/1711.07767)
1052 | - github: [https://github.com/ruinmessi/RFBNet](https://github.com/ruinmessi/RFBNet)
1053 | 
1054 | **Deep Feature Pyramid Reconfiguration for Object Detection**
1055 | 
1056 | - intro: ECCV 2018
1057 | - arXiv: https://arxiv.org/abs/1808.07993
1058 | 
1059 | **Unsupervised Hard Example Mining from Videos for Improved Object Detection**
1060 | 
1061 | - intro: ECCV 2018
1062 | - arXiv: https://arxiv.org/abs/1808.04285
1063 | 
1064 | **Acquisition of Localization Confidence for Accurate Object Detection**
1065 | 
1066 | - intro: ECCV 2018
1067 | - arXiv: https://arxiv.org/abs/1807.11590
1068 | - github: https://github.com/vacancy/PreciseRoIPooling
1069 | 
1070 | **Toward Scale-Invariance and Position-Sensitive Region Proposal Networks**
1071 | 
1072 | - intro: ECCV 2018
1073 | - arXiv: https://arxiv.org/abs/1807.09528
1074 | 
1075 | **MetaAnchor: Learning to Detect Objects with Customized Anchors**
1076 | 
1077 | - arxiv: https://arxiv.org/abs/1807.00980
1078 | 
1079 | **Relation Network for Object Detection**
1080 | 
1081 | - intro: CVPR 2018
1082 | - arxiv: https://arxiv.org/abs/1711.11575
1083 | - github:https://github.com/msracver/Relation-Networks-for-Object-Detection
1084 | 
1085 | **Quantization Mimic: Towards Very Tiny CNN for Object Detection**
1086 | 
1087 | - Tsinghua University1 & The Chinese University of Hong Kong2 &SenseTime3
1088 | - arxiv: https://arxiv.org/abs/1805.02152
1089 | 
1090 | **Learning Rich Features for Image Manipulation Detection**
1091 | 
1092 | - intro: CVPR 2018 Camera Ready
1093 | - arxiv: https://arxiv.org/abs/1805.04953
1094 | 
1095 | **SNIPER: Efficient Multi-Scale Training**
1096 | 
1097 | - arxiv:https://arxiv.org/abs/1805.09300
1098 | - github:https://github.com/mahyarnajibi/SNIPER
1099 | 
1100 | **Soft Sampling for Robust Object Detection**
1101 | 
1102 | - intro: the robustness of object detection under the presence of missing annotations
1103 | - arxiv:https://arxiv.org/abs/1806.06986
1104 | 
1105 | **Cost-effective Object Detection: Active Sample Mining with Switchable Selection Criteria**
1106 | 
1107 | - intro: TNNLS 2018
1108 | - arxiv:https://arxiv.org/abs/1807.00147
1109 | - code: http://kezewang.com/codes/ASM_ver1.zip
1110 | 
1111 | ## Other
1112 | 
1113 | **R3-Net: A Deep Network for Multi-oriented Vehicle Detection in Aerial Images and Videos**
1114 | 
1115 | - arxiv: https://arxiv.org/abs/1808.05560
1116 | - youtube: https://youtu.be/xCYD-tYudN0
1117 | 
1118 | # Detection Toolbox
1119 | 
1120 | - [Detectron(FAIR)](https://github.com/facebookresearch/Detectron): Detectron is Facebook AI Research's software system that implements state-of-the-art object detection algorithms, including [Mask R-CNN](https://arxiv.org/abs/1703.06870). It is written in Python and powered by the [Caffe2](https://github.com/caffe2/caffe2) deep learning framework.
1121 | - [Detectron2](https://github.com/facebookresearch/detectron2): Detectron2 is FAIR's next-generation research platform for object detection and segmentation.
1122 | - [maskrcnn-benchmark(FAIR)](https://github.com/facebookresearch/maskrcnn-benchmark): Fast, modular reference implementation of Instance Segmentation and Object Detection algorithms in PyTorch.
1123 | - [mmdetection(SenseTime&CUHK)](https://github.com/open-mmlab/mmdetection): mmdetection is an open source object detection toolbox based on PyTorch. It is a part of the open-mmlab project developed by [Multimedia Laboratory, CUHK](http://mmlab.ie.cuhk.edu.hk/).
1124 | 


--------------------------------------------------------------------------------