├── .gitignore ├── LICENSE ├── README.md ├── examples ├── maskrcnn.py ├── maskrcnn_requirements.py ├── yolov5.py └── yolov5_requirements.py ├── pytorch_clip_bbox ├── __init__.py └── clip_bbox.py ├── requirements.txt ├── resources └── preds.jpg ├── setup.cfg └── setup.py /.gitignore: -------------------------------------------------------------------------------- 1 | # Byte-compiled / optimized / DLL files 2 | __pycache__/ 3 | *.py[cod] 4 | *$py.class 5 | 6 | # C extensions 7 | *.so 8 | 9 | # Distribution / packaging 10 | .Python 11 | build/ 12 | develop-eggs/ 13 | dist/ 14 | downloads/ 15 | eggs/ 16 | .eggs/ 17 | lib/ 18 | lib64/ 19 | parts/ 20 | sdist/ 21 | var/ 22 | wheels/ 23 | pip-wheel-metadata/ 24 | share/python-wheels/ 25 | *.egg-info/ 26 | .installed.cfg 27 | *.egg 28 | MANIFEST 29 | 30 | # PyInstaller 31 | # Usually these files are written by a python script from a template 32 | # before PyInstaller builds the exe, so as to inject date/other infos into it. 33 | *.manifest 34 | *.spec 35 | 36 | # Installer logs 37 | pip-log.txt 38 | pip-delete-this-directory.txt 39 | 40 | # Unit test / coverage reports 41 | htmlcov/ 42 | .tox/ 43 | .nox/ 44 | .coverage 45 | .coverage.* 46 | .cache 47 | nosetests.xml 48 | coverage.xml 49 | *.cover 50 | *.py,cover 51 | .hypothesis/ 52 | .pytest_cache/ 53 | 54 | # Translations 55 | *.mo 56 | *.pot 57 | 58 | # Django stuff: 59 | *.log 60 | local_settings.py 61 | db.sqlite3 62 | db.sqlite3-journal 63 | 64 | # Flask stuff: 65 | instance/ 66 | .webassets-cache 67 | 68 | # Scrapy stuff: 69 | .scrapy 70 | 71 | # Sphinx documentation 72 | docs/_build/ 73 | 74 | # PyBuilder 75 | target/ 76 | 77 | # Jupyter Notebook 78 | .ipynb_checkpoints 79 | 80 | # IPython 81 | profile_default/ 82 | ipython_config.py 83 | 84 | # pyenv 85 | .python-version 86 | 87 | # pipenv 88 | # According to pypa/pipenv#598, it is recommended to include Pipfile.lock in version control. 89 | # However, in case of collaboration, if having platform-specific dependencies or dependencies 90 | # having no cross-platform support, pipenv may install dependencies that don't work, or not 91 | # install all needed dependencies. 92 | #Pipfile.lock 93 | 94 | # PEP 582; used by e.g. github.com/David-OConnor/pyflow 95 | __pypackages__/ 96 | 97 | # Celery stuff 98 | celerybeat-schedule 99 | celerybeat.pid 100 | 101 | # SageMath parsed files 102 | *.sage.py 103 | 104 | # Environments 105 | .env 106 | .venv 107 | env/ 108 | venv/ 109 | ENV/ 110 | env.bak/ 111 | venv.bak/ 112 | 113 | # Spyder project settings 114 | .spyderproject 115 | .spyproject 116 | 117 | # Rope project settings 118 | .ropeproject 119 | 120 | # mkdocs documentation 121 | /site 122 | 123 | # mypy 124 | .mypy_cache/ 125 | .dmypy.json 126 | dmypy.json 127 | 128 | # Pyre type checker 129 | .pyre/ 130 | -------------------------------------------------------------------------------- /LICENSE: -------------------------------------------------------------------------------- 1 | Apache License 2 | Version 2.0, January 2004 3 | http://www.apache.org/licenses/ 4 | 5 | TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION 6 | 7 | 1. Definitions. 8 | 9 | "License" shall mean the terms and conditions for use, reproduction, 10 | and distribution as defined by Sections 1 through 9 of this document. 11 | 12 | "Licensor" shall mean the copyright owner or entity authorized by 13 | the copyright owner that is granting the License. 14 | 15 | "Legal Entity" shall mean the union of the acting entity and all 16 | other entities that control, are controlled by, or are under common 17 | control with that entity. For the purposes of this definition, 18 | "control" means (i) the power, direct or indirect, to cause the 19 | direction or management of such entity, whether by contract or 20 | otherwise, or (ii) ownership of fifty percent (50%) or more of the 21 | outstanding shares, or (iii) beneficial ownership of such entity. 22 | 23 | "You" (or "Your") shall mean an individual or Legal Entity 24 | exercising permissions granted by this License. 25 | 26 | "Source" form shall mean the preferred form for making modifications, 27 | including but not limited to software source code, documentation 28 | source, and configuration files. 29 | 30 | "Object" form shall mean any form resulting from mechanical 31 | transformation or translation of a Source form, including but 32 | not limited to compiled object code, generated documentation, 33 | and conversions to other media types. 34 | 35 | "Work" shall mean the work of authorship, whether in Source or 36 | Object form, made available under the License, as indicated by a 37 | copyright notice that is included in or attached to the work 38 | (an example is provided in the Appendix below). 39 | 40 | "Derivative Works" shall mean any work, whether in Source or Object 41 | form, that is based on (or derived from) the Work and for which the 42 | editorial revisions, annotations, elaborations, or other modifications 43 | represent, as a whole, an original work of authorship. For the purposes 44 | of this License, Derivative Works shall not include works that remain 45 | separable from, or merely link (or bind by name) to the interfaces of, 46 | the Work and Derivative Works thereof. 47 | 48 | "Contribution" shall mean any work of authorship, including 49 | the original version of the Work and any modifications or additions 50 | to that Work or Derivative Works thereof, that is intentionally 51 | submitted to Licensor for inclusion in the Work by the copyright owner 52 | or by an individual or Legal Entity authorized to submit on behalf of 53 | the copyright owner. For the purposes of this definition, "submitted" 54 | means any form of electronic, verbal, or written communication sent 55 | to the Licensor or its representatives, including but not limited to 56 | communication on electronic mailing lists, source code control systems, 57 | and issue tracking systems that are managed by, or on behalf of, the 58 | Licensor for the purpose of discussing and improving the Work, but 59 | excluding communication that is conspicuously marked or otherwise 60 | designated in writing by the copyright owner as "Not a Contribution." 61 | 62 | "Contributor" shall mean Licensor and any individual or Legal Entity 63 | on behalf of whom a Contribution has been received by Licensor and 64 | subsequently incorporated within the Work. 65 | 66 | 2. Grant of Copyright License. Subject to the terms and conditions of 67 | this License, each Contributor hereby grants to You a perpetual, 68 | worldwide, non-exclusive, no-charge, royalty-free, irrevocable 69 | copyright license to reproduce, prepare Derivative Works of, 70 | publicly display, publicly perform, sublicense, and distribute the 71 | Work and such Derivative Works in Source or Object form. 72 | 73 | 3. Grant of Patent License. Subject to the terms and conditions of 74 | this License, each Contributor hereby grants to You a perpetual, 75 | worldwide, non-exclusive, no-charge, royalty-free, irrevocable 76 | (except as stated in this section) patent license to make, have made, 77 | use, offer to sell, sell, import, and otherwise transfer the Work, 78 | where such license applies only to those patent claims licensable 79 | by such Contributor that are necessarily infringed by their 80 | Contribution(s) alone or by combination of their Contribution(s) 81 | with the Work to which such Contribution(s) was submitted. If You 82 | institute patent litigation against any entity (including a 83 | cross-claim or counterclaim in a lawsuit) alleging that the Work 84 | or a Contribution incorporated within the Work constitutes direct 85 | or contributory patent infringement, then any patent licenses 86 | granted to You under this License for that Work shall terminate 87 | as of the date such litigation is filed. 88 | 89 | 4. Redistribution. You may reproduce and distribute copies of the 90 | Work or Derivative Works thereof in any medium, with or without 91 | modifications, and in Source or Object form, provided that You 92 | meet the following conditions: 93 | 94 | (a) You must give any other recipients of the Work or 95 | Derivative Works a copy of this License; and 96 | 97 | (b) You must cause any modified files to carry prominent notices 98 | stating that You changed the files; and 99 | 100 | (c) You must retain, in the Source form of any Derivative Works 101 | that You distribute, all copyright, patent, trademark, and 102 | attribution notices from the Source form of the Work, 103 | excluding those notices that do not pertain to any part of 104 | the Derivative Works; and 105 | 106 | (d) If the Work includes a "NOTICE" text file as part of its 107 | distribution, then any Derivative Works that You distribute must 108 | include a readable copy of the attribution notices contained 109 | within such NOTICE file, excluding those notices that do not 110 | pertain to any part of the Derivative Works, in at least one 111 | of the following places: within a NOTICE text file distributed 112 | as part of the Derivative Works; within the Source form or 113 | documentation, if provided along with the Derivative Works; or, 114 | within a display generated by the Derivative Works, if and 115 | wherever such third-party notices normally appear. The contents 116 | of the NOTICE file are for informational purposes only and 117 | do not modify the License. You may add Your own attribution 118 | notices within Derivative Works that You distribute, alongside 119 | or as an addendum to the NOTICE text from the Work, provided 120 | that such additional attribution notices cannot be construed 121 | as modifying the License. 122 | 123 | You may add Your own copyright statement to Your modifications and 124 | may provide additional or different license terms and conditions 125 | for use, reproduction, or distribution of Your modifications, or 126 | for any such Derivative Works as a whole, provided Your use, 127 | reproduction, and distribution of the Work otherwise complies with 128 | the conditions stated in this License. 129 | 130 | 5. Submission of Contributions. Unless You explicitly state otherwise, 131 | any Contribution intentionally submitted for inclusion in the Work 132 | by You to the Licensor shall be under the terms and conditions of 133 | this License, without any additional terms or conditions. 134 | Notwithstanding the above, nothing herein shall supersede or modify 135 | the terms of any separate license agreement you may have executed 136 | with Licensor regarding such Contributions. 137 | 138 | 6. Trademarks. This License does not grant permission to use the trade 139 | names, trademarks, service marks, or product names of the Licensor, 140 | except as required for reasonable and customary use in describing the 141 | origin of the Work and reproducing the content of the NOTICE file. 142 | 143 | 7. Disclaimer of Warranty. Unless required by applicable law or 144 | agreed to in writing, Licensor provides the Work (and each 145 | Contributor provides its Contributions) on an "AS IS" BASIS, 146 | WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or 147 | implied, including, without limitation, any warranties or conditions 148 | of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A 149 | PARTICULAR PURPOSE. You are solely responsible for determining the 150 | appropriateness of using or redistributing the Work and assume any 151 | risks associated with Your exercise of permissions under this License. 152 | 153 | 8. Limitation of Liability. In no event and under no legal theory, 154 | whether in tort (including negligence), contract, or otherwise, 155 | unless required by applicable law (such as deliberate and grossly 156 | negligent acts) or agreed to in writing, shall any Contributor be 157 | liable to You for damages, including any direct, indirect, special, 158 | incidental, or consequential damages of any character arising as a 159 | result of this License or out of the use or inability to use the 160 | Work (including but not limited to damages for loss of goodwill, 161 | work stoppage, computer failure or malfunction, or any and all 162 | other commercial damages or losses), even if such Contributor 163 | has been advised of the possibility of such damages. 164 | 165 | 9. Accepting Warranty or Additional Liability. While redistributing 166 | the Work or Derivative Works thereof, You may choose to offer, 167 | and charge a fee for, acceptance of support, warranty, indemnity, 168 | or other liability obligations and/or rights consistent with this 169 | License. However, in accepting such obligations, You may act only 170 | on Your own behalf and on Your sole responsibility, not on behalf 171 | of any other Contributor, and only if You agree to indemnify, 172 | defend, and hold each Contributor harmless for any liability 173 | incurred by, or claims asserted against, such Contributor by reason 174 | of your accepting any such warranty or additional liability. 175 | 176 | END OF TERMS AND CONDITIONS 177 | 178 | APPENDIX: How to apply the Apache License to your work. 179 | 180 | To apply the Apache License to your work, attach the following 181 | boilerplate notice, with the fields enclosed by brackets "[]" 182 | replaced with your own identifying information. (Don't include 183 | the brackets!) The text should be enclosed in the appropriate 184 | comment syntax for the file format. We also recommend that a 185 | file or class name and description of purpose be included on the 186 | same "printed page" as the copyright notice for easier 187 | identification within third-party archives. 188 | 189 | Copyright [yyyy] [name of copyright owner] 190 | 191 | Licensed under the Apache License, Version 2.0 (the "License"); 192 | you may not use this file except in compliance with the License. 193 | You may obtain a copy of the License at 194 | 195 | http://www.apache.org/licenses/LICENSE-2.0 196 | 197 | Unless required by applicable law or agreed to in writing, software 198 | distributed under the License is distributed on an "AS IS" BASIS, 199 | WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 200 | See the License for the specific language governing permissions and 201 | limitations under the License. 202 | -------------------------------------------------------------------------------- /README.md: -------------------------------------------------------------------------------- 1 | # pytorch_clip_bbox: Implementation of the CLIP guided bbox ranking for Object Detection. 2 | 3 |

4 | 5 |

6 | 7 | Pytorch based library to rank predicted bounding boxes using text/image user's prompts. 8 | 9 | Usually, object detection models trains to detect common classes of objects such as "car", "person", "cup", "bottle". 10 | But sometimes we need to detect more complex classes such as "lady in the red dress", "bottle of whiskey", or "where is my red cup" instead of "person", "bottle", "cup" respectively. 11 | One way to solve this problem is to train more complex detectors that can detect more complex classes, 12 | but we propose to use text-driven object detection that allows detecting any complex classes that can be described by natural language. 13 | This library is written to rank predicted bounding boxes using text/image descriptions of complex classes. 14 | 15 | ## Install package 16 | 17 | ```bash 18 | pip install pytorch_clip_bbox 19 | ``` 20 | 21 | ## Install the latest version 22 | 23 | ```bash 24 | pip install --upgrade git+https://github.com/bes-dev/pytorch_clip_bbox.git 25 | ``` 26 | 27 | ## Features 28 | - The library supports multiple prompts (images or texts) as targets for filtering. 29 | - The library automatically detects the language of the input text, and multilingual translate it via google translate. 30 | - The library supports the original CLIP model by OpenAI and ruCLIP model by SberAI. 31 | - Simple integration with different object detection models. 32 | 33 | ## Usage 34 | 35 | We provide examples to integrate our library with different popular object detectors like: [YOLOv5](examples/yolov5.py), [MaskRCNN](examples/maskrcnn.py). 36 | Please, follow to [examples](examples/) to find more examples. 37 | 38 | ### Simple example to integrate pytorch_clip_bbox with MaskRCNN model 39 | 40 | ```bash 41 | $ pip install -r wheel cython opencv-python numpy torch torchvision pytorch_clip_bbox 42 | ``` 43 | 44 | ```python 45 | import argparse 46 | import random 47 | import cv2 48 | import numpy as np 49 | import torch 50 | import torchvision.transforms as T 51 | import torchvision 52 | from pytorch_clip_bbox import ClipBBOX 53 | 54 | def get_coloured_mask(mask): 55 | colours = [[0, 255, 0],[0, 0, 255],[255, 0, 0],[0, 255, 255],[255, 255, 0],[255, 0, 255],[80, 70, 180],[250, 80, 190],[245, 145, 50],[70, 150, 250],[50, 190, 190]] 56 | r = np.zeros_like(mask).astype(np.uint8) 57 | g = np.zeros_like(mask).astype(np.uint8) 58 | b = np.zeros_like(mask).astype(np.uint8) 59 | c = colours[random.randrange(0,10)] 60 | r[mask == 1], g[mask == 1], b[mask == 1] = c 61 | coloured_mask = np.stack([r, g, b], axis=2) 62 | return coloured_mask, c 63 | 64 | def main(args): 65 | # build detector 66 | detector = torchvision.models.detection.maskrcnn_resnet50_fpn(pretrained=True).eval().to(args.device) 67 | clip_bbox = ClipBBOX(clip_type=args.clip_type).to(args.device) 68 | # add prompts 69 | if args.text_prompt is not None: 70 | for prompt in args.text_prompt.split(","): 71 | clip_bbox.add_prompt(text=prompt) 72 | if args.image_prompt is not None: 73 | image = cv2.cvtColor(cv2.imread(args.image_prompt), cv2.COLOR_BGR2RGB) 74 | image = torch.from_numpy(image).permute(2, 0, 1).unsqueeze(0) 75 | image = img / 255.0 76 | clip_bbox.add_prompt(image=image) 77 | image = cv2.imread(args.image) 78 | pred = detector([ 79 | T.ToTensor()(cv2.cvtColor(image, cv2.COLOR_BGR2RGB)).to(args.device) 80 | ]) 81 | pred_score = list(pred[0]['scores'].detach().cpu().numpy()) 82 | pred_threshold = [pred_score.index(x) for x in pred_score if x > args.confidence][-1] 83 | boxes = [[int(b) for b in box] for box in list(pred[0]['boxes'].detach().cpu().numpy())][:pred_threshold + 1] 84 | masks = (pred[0]['masks'] > 0.5).squeeze().detach().cpu().numpy()[:pred_threshold + 1] 85 | ranking = clip_bbox(image, boxes, top_k=args.top_k) 86 | for key in ranking.keys(): 87 | if key == "loss": 88 | continue 89 | for box in ranking[key]["ranking"]: 90 | mask, color = get_coloured_mask(masks[box["idx"]]) 91 | image = cv2.addWeighted(image, 1, mask, 0.5, 0) 92 | x1, y1, x2, y2 = box["rect"] 93 | cv2.rectangle(image, (x1, y1), (x2, y2), color, 6) 94 | cv2.rectangle(image, (x1, y1), (x2, y1-100), color, -1) 95 | cv2.putText(image, ranking[key]["src"], (x1 + 5, y1 - 10), cv2.FONT_HERSHEY_SIMPLEX, 4, (0, 0, 0), thickness=5) 96 | if args.output_image is None: 97 | cv2.imshow("image", image) 98 | cv2.waitKey() 99 | else: 100 | cv2.imwrite(args.output_image, image) 101 | 102 | 103 | if __name__ == "__main__": 104 | parser = argparse.ArgumentParser() 105 | parser.add_argument("-i", "--image", type=str, help="Input image.") 106 | parser.add_argument("--device", type=str, default="cuda:0", help="inference device.") 107 | parser.add_argument("--confidence", type=float, default=0.7, help="confidence threshold [MaskRCNN].") 108 | parser.add_argument("--text-prompt", type=str, default=None, help="Text prompt.") 109 | parser.add_argument("--image-prompt", type=str, default=None, help="Image prompt.") 110 | parser.add_argument("--clip-type", type=str, default="clip_vit_b32", help="Type of CLIP model [ruclip, clip_vit_b32, clip_vit_b16].") 111 | parser.add_argument("--top-k", type=int, default=1, help="top_k predictions will be returned.") 112 | parser.add_argument("--output-image", type=str, default=None, help="Output image name.") 113 | args = parser.parse_args() 114 | main(args) 115 | ``` -------------------------------------------------------------------------------- /examples/maskrcnn.py: -------------------------------------------------------------------------------- 1 | """ 2 | Copyright 2021 by Sergei Belousov 3 | Licensed under the Apache License, Version 2.0 (the "License"); 4 | you may not use this file except in compliance with the License. 5 | You may obtain a copy of the License at 6 | http://www.apache.org/licenses/LICENSE-2.0 7 | Unless required by applicable law or agreed to in writing, software 8 | distributed under the License is distributed on an "AS IS" BASIS, 9 | WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 10 | See the License for the specific language governing permissions and 11 | limitations under the License. 12 | """ 13 | import argparse 14 | import random 15 | import cv2 16 | import numpy as np 17 | import torch 18 | import torchvision.transforms as T 19 | import torchvision 20 | from pytorch_clip_bbox import ClipBBOX 21 | 22 | def get_coloured_mask(mask): 23 | colours = [[0, 255, 0],[0, 0, 255],[255, 0, 0],[0, 255, 255],[255, 255, 0],[255, 0, 255],[80, 70, 180],[250, 80, 190],[245, 145, 50],[70, 150, 250],[50, 190, 190]] 24 | r = np.zeros_like(mask).astype(np.uint8) 25 | g = np.zeros_like(mask).astype(np.uint8) 26 | b = np.zeros_like(mask).astype(np.uint8) 27 | c = colours[random.randrange(0,10)] 28 | r[mask == 1], g[mask == 1], b[mask == 1] = c 29 | coloured_mask = np.stack([r, g, b], axis=2) 30 | return coloured_mask, c 31 | 32 | def main(args): 33 | # build detector 34 | detector = torchvision.models.detection.maskrcnn_resnet50_fpn(pretrained=True).eval().to(args.device) 35 | clip_bbox = ClipBBOX(clip_type=args.clip_type).to(args.device) 36 | # add prompts 37 | if args.text_prompt is not None: 38 | for prompt in args.text_prompt.split(","): 39 | clip_bbox.add_prompt(text=prompt) 40 | if args.image_prompt is not None: 41 | image = cv2.cvtColor(cv2.imread(args.image_prompt), cv2.COLOR_BGR2RGB) 42 | image = torch.from_numpy(image).permute(2, 0, 1).unsqueeze(0) 43 | image = img / 255.0 44 | clip_bbox.add_prompt(image=image) 45 | image = cv2.imread(args.image) 46 | pred = detector([ 47 | T.ToTensor()(cv2.cvtColor(image, cv2.COLOR_BGR2RGB)).to(args.device) 48 | ]) 49 | pred_score = list(pred[0]['scores'].detach().cpu().numpy()) 50 | pred_threshold = [pred_score.index(x) for x in pred_score if x > args.confidence][-1] 51 | boxes = [[int(b) for b in box] for box in list(pred[0]['boxes'].detach().cpu().numpy())][:pred_threshold + 1] 52 | masks = (pred[0]['masks'] > 0.5).squeeze().detach().cpu().numpy()[:pred_threshold + 1] 53 | ranking = clip_bbox(image, boxes, top_k=args.top_k) 54 | for key in ranking.keys(): 55 | if key == "loss": 56 | continue 57 | for box in ranking[key]["ranking"]: 58 | mask, color = get_coloured_mask(masks[box["idx"]]) 59 | image = cv2.addWeighted(image, 1, mask, 0.5, 0) 60 | x1, y1, x2, y2 = box["rect"] 61 | cv2.rectangle(image, (x1, y1), (x2, y2), color, 6) 62 | cv2.rectangle(image, (x1, y1), (x2, y1-100), color, -1) 63 | cv2.putText(image, ranking[key]["src"], (x1 + 5, y1 - 10), cv2.FONT_HERSHEY_SIMPLEX, 4, (0, 0, 0), thickness=5) 64 | if args.output_image is None: 65 | cv2.imshow("image", image) 66 | cv2.waitKey() 67 | else: 68 | cv2.imwrite(args.output_image, image) 69 | 70 | 71 | if __name__ == "__main__": 72 | parser = argparse.ArgumentParser() 73 | parser.add_argument("-i", "--image", type=str, help="Input image.") 74 | parser.add_argument("--device", type=str, default="cuda:0", help="inference device.") 75 | parser.add_argument("--confidence", type=float, default=0.7, help="confidence threshold [MaskRCNN].") 76 | parser.add_argument("--text-prompt", type=str, default=None, help="Text prompt.") 77 | parser.add_argument("--image-prompt", type=str, default=None, help="Image prompt.") 78 | parser.add_argument("--clip-type", type=str, default="clip_vit_b32", help="Type of CLIP model [ruclip, clip_vit_b32, clip_vit_b16].") 79 | parser.add_argument("--top-k", type=int, default=1, help="top_k predictions will be returned.") 80 | parser.add_argument("--output-image", type=str, default=None, help="Output image name.") 81 | args = parser.parse_args() 82 | main(args) 83 | -------------------------------------------------------------------------------- /examples/maskrcnn_requirements.py: -------------------------------------------------------------------------------- 1 | wheel 2 | cython 3 | opencv-python 4 | numpy 5 | torch 6 | torchvision 7 | pytorch_clip_bbox 8 | -------------------------------------------------------------------------------- /examples/yolov5.py: -------------------------------------------------------------------------------- 1 | """ 2 | Copyright 2021 by Sergei Belousov 3 | Licensed under the Apache License, Version 2.0 (the "License"); 4 | you may not use this file except in compliance with the License. 5 | You may obtain a copy of the License at 6 | http://www.apache.org/licenses/LICENSE-2.0 7 | Unless required by applicable law or agreed to in writing, software 8 | distributed under the License is distributed on an "AS IS" BASIS, 9 | WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 10 | See the License for the specific language governing permissions and 11 | limitations under the License. 12 | """ 13 | import argparse 14 | import cv2 15 | import torch 16 | from pytorch_clip_bbox import ClipBBOX 17 | 18 | def extract_boxes(detections): 19 | boxes = [] 20 | for i in range(detections.xyxy[0].size(0)): 21 | x1, y1, x2, y2, confidence, idx = detections.xyxy[0][i] 22 | boxes.append([int(x1), int(y1), int(x2), int(y2)]) 23 | return boxes 24 | 25 | def main(args): 26 | # build detector 27 | detector = torch.hub.load("ultralytics/yolov5", "yolov5s").to(args.device) 28 | clip_bbox = ClipBBOX(clip_type=args.clip_type).to(args.device) 29 | # add prompts 30 | if args.text_prompt is not None: 31 | for prompt in args.text_prompt.split(","): 32 | clip_bbox.add_prompt(text=prompt) 33 | if args.image_prompt is not None: 34 | image = cv2.cvtColor(cv2.imread(args.image_prompt), cv2.COLOR_BGR2RGB) 35 | image = torch.from_numpy(image).permute(2, 0, 1).unsqueeze(0) 36 | image = img / 255.0 37 | clip_bbox.add_prompt(image=image) 38 | image = cv2.imread(args.image) 39 | detections = detector(cv2.cvtColor(image, cv2.COLOR_BGR2RGB)) 40 | boxes = extract_boxes(detections) 41 | ranking = clip_bbox(image, boxes, top_k=args.top_k) 42 | for key in ranking.keys(): 43 | if key == "loss": 44 | continue 45 | for box in ranking[key]["ranking"]: 46 | x1, y1, x2, y2 = box["rect"] 47 | cv2.rectangle(image, (x1, y1), (x2, y2), (0, 255, 0), 6) 48 | cv2.rectangle(image, (x1, y1), (x2, y1-100), (0, 255, 0), -1) 49 | cv2.putText(image, ranking[key]["src"], (x1 + 5, y1 - 10), cv2.FONT_HERSHEY_SIMPLEX, 4, (0, 0, 0), thickness=5) 50 | if args.output_image is None: 51 | cv2.imshow("image", image) 52 | cv2.waitKey() 53 | else: 54 | cv2.imwrite(args.output_image, image) 55 | 56 | 57 | if __name__ == "__main__": 58 | parser = argparse.ArgumentParser() 59 | parser.add_argument("-i", "--image", type=str, help="Input image.") 60 | parser.add_argument("--device", type=str, default="cuda:0", help="inference device.") 61 | parser.add_argument("--text-prompt", type=str, default=None, help="Text prompt.") 62 | parser.add_argument("--image-prompt", type=str, default=None, help="Image prompt.") 63 | parser.add_argument("--clip-type", type=str, default="clip_vit_b32", help="Type of CLIP model [ruclip, clip_vit_b32, clip_vit_b16].") 64 | parser.add_argument("--top-k", type=int, default=1, help="top_k predictions will be returned.") 65 | parser.add_argument("--output-image", type=str, default=None, help="Output image name.") 66 | args = parser.parse_args() 67 | main(args) 68 | -------------------------------------------------------------------------------- /examples/yolov5_requirements.py: -------------------------------------------------------------------------------- 1 | pytorch_clip_bbox 2 | matplotlib>=3.2.2 3 | numpy>=1.18.5 4 | opencv-python>=4.1.2 5 | Pillow>=7.1.2 6 | PyYAML>=5.3.1 7 | requests>=2.23.0 8 | scipy>=1.4.1 9 | torch>=1.7.0 10 | torchvision>=0.8.1 11 | tqdm>=4.41.0 12 | tensorboard>=2.4.1 13 | pandas>=1.1.4 14 | seaborn>=0.11.0 15 | thop 16 | -------------------------------------------------------------------------------- /pytorch_clip_bbox/__init__.py: -------------------------------------------------------------------------------- 1 | """ 2 | Copyright 2021 by Sergei Belousov 3 | Licensed under the Apache License, Version 2.0 (the "License"); 4 | you may not use this file except in compliance with the License. 5 | You may obtain a copy of the License at 6 | http://www.apache.org/licenses/LICENSE-2.0 7 | Unless required by applicable law or agreed to in writing, software 8 | distributed under the License is distributed on an "AS IS" BASIS, 9 | WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 10 | See the License for the specific language governing permissions and 11 | limitations under the License. 12 | """ 13 | from .clip_bbox import ClipBBOX 14 | -------------------------------------------------------------------------------- /pytorch_clip_bbox/clip_bbox.py: -------------------------------------------------------------------------------- 1 | """ 2 | Copyright 2021 by Sergei Belousov 3 | Licensed under the Apache License, Version 2.0 (the "License"); 4 | you may not use this file except in compliance with the License. 5 | You may obtain a copy of the License at 6 | http://www.apache.org/licenses/LICENSE-2.0 7 | Unless required by applicable law or agreed to in writing, software 8 | distributed under the License is distributed on an "AS IS" BASIS, 9 | WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 10 | See the License for the specific language governing permissions and 11 | limitations under the License. 12 | """ 13 | from typing import List, Dict, Tuple, Optional 14 | import cv2 15 | import numpy as np 16 | import torch 17 | import torch.nn as nn 18 | from pytorch_clip_guided_loss import get_clip_guided_loss 19 | from pytorch_clip_guided_loss.clip_guided_loss import CLIPPrompt 20 | 21 | 22 | class ClipBBOX(nn.Module): 23 | """ Implementation of the CLIP guided bbox refinement for Object Detection. 24 | Arguments: 25 | clip_type (str): type of the CLIP model. 26 | """ 27 | def __init__( 28 | self, 29 | clip_type: str = "clip_vit_b32" 30 | ): 31 | super().__init__() 32 | # CLIP guided loss 33 | self.clip_loss = get_clip_guided_loss(clip_type, input_range=(0.0, 1.0)) 34 | self.input_size = self.clip_loss.image_processor[0].size 35 | # utils 36 | self.register_buffer("device_info", torch.tensor(0)) 37 | 38 | def add_prompt( 39 | self, 40 | image: Optional[torch.Tensor] = None, 41 | text: Optional[str] = None, 42 | weight: float = 1.0, 43 | label: Optional[str] = None, 44 | store_src: bool = True 45 | ) -> str: 46 | """Add prompt to loss function. 47 | Arguments: 48 | image (torch.Tensor): input image [Optional]. 49 | text (str): input text [Optional]. 50 | weight (float): importance of the prompt. 51 | label (str): label of the prompt [Optional]. 52 | store_src (bool): store source data of the prompt. 53 | Returns: 54 | label (src): label of the prompt. 55 | """ 56 | return self.clip_loss.add_prompt(image, text, weight, label, store_src) 57 | 58 | def get_prompt(self, label: str) -> Optional[CLIPPrompt]: 59 | """Get prompt if available. 60 | Arguments: 61 | label (str): label of the prompt [Optional]. 62 | Returns: 63 | prompt (CLIPPrompt or None): prompt [Optional]. 64 | """ 65 | return self.clip_loss.get_prompt(label) 66 | 67 | def get_prompts_list(self) -> List[str]: 68 | """Get list of all available prompts. 69 | Returns: 70 | prompts (list): list of prompts labels. 71 | """ 72 | return self.clip_loss.get_prompts_list() 73 | 74 | def delete_prompt(self, label: Optional[str] = None) -> None: 75 | """Add prompt to loss function. 76 | Arguments: 77 | label (str): label of the prompt to delete [Optional]. 78 | """ 79 | return self.clip_loss.delete_prompt(label) 80 | 81 | def clear_prompts(self) -> None: 82 | """Delete all available prompts.""" 83 | return self.clip_loss.clear_prompts() 84 | 85 | @torch.no_grad() 86 | def forward( 87 | self, 88 | img: np.array, 89 | boxes: List[Tuple[int, int, int, int]], 90 | is_rgb: bool = False, 91 | top_k: int = 1, 92 | batch_size: int = 128 93 | ): 94 | """ CLIP guided filter for input bounding boxes. 95 | Argument: 96 | img (np.array): input image. 97 | boxes (List[Tuple[int, int, int, int]]): input bounding boxes in format [(xmin, ymin, xmax, ymax)] 98 | is_rgb (bool): is imput image in RGB/BGR format. 99 | top_k (int): top k best matches will be returned. 100 | Use top_k = -1 to return all boxes in ranked order. 101 | batch_size (int): batch size. 102 | Returns: 103 | outputs (List[Dict]): predicts in format: 104 | [{"rect": [x, y, w, h], "loss": loss_val}] 105 | """ 106 | _img = img.copy() if is_rgb else cv2.cvtColor(img, cv2.COLOR_BGR2RGB) 107 | batch = self._prepare_batch(_img, boxes).to(self.device_info.device) 108 | loss = self._predict_clip_loss(batch, batch_size) 109 | outputs = self._generate_output(boxes, loss, top_k) 110 | return outputs 111 | 112 | def _prepare_batch(self, img: np.array, boxes: List[Tuple[int, int, int, int]]) -> torch.Tensor: 113 | """ Crop region proposals and generate batch 114 | Argument: 115 | img (np.array): input image in opencv format (H, W, 3). 116 | boxes (List[Tuple[int, int, int, int]]): input boxes in format [(xmin, ymin, xmax, ymax)] 117 | Returns: 118 | batch (torch.Tensor): output batch (B, C, H, W). 119 | """ 120 | batch = [] 121 | for x1, y1, x2, y2 in boxes: 122 | crop = cv2.resize(img[y1:y2, x1:x2], self.input_size) 123 | batch.append(torch.from_numpy(crop).permute(2, 0, 1).unsqueeze(0)) 124 | batch = torch.cat(batch, dim=0) 125 | # normalize batch 126 | batch = batch / (batch.max() + 1e-8) 127 | return batch 128 | 129 | def _predict_clip_loss( 130 | self, 131 | batch_full: torch.Tensor, 132 | batch_size: int = 128 133 | ) -> torch.Tensor: 134 | """ Predict CLIP loss for region proposals using user's prompts. 135 | Argument: 136 | batch_full (torch.Tensor): input batch (B, C, H, W). 137 | prompt_label (str): prompt label that uses for ranking. 138 | batch_size (int): batch size. 139 | Returns: 140 | loss (Dict[str, torch.Tensor]): output batch {"prompt_label": (B, )}. 141 | """ 142 | loss = {} 143 | id_start = 0 144 | while id_start < batch_full.size(0): 145 | id_stop = min(id_start + batch_size, batch_full.size(0)) 146 | batch = batch_full[id_start:id_stop] 147 | predicted_loss = self.clip_loss.image_loss(image=batch, reduce=None) 148 | for key, val in predicted_loss.items(): 149 | if not key in loss: 150 | loss[key] = [] 151 | loss[key].append(val.cpu()) 152 | id_start = id_stop 153 | for key, val in loss.items(): 154 | loss[key] = torch.cat(val, dim=0) 155 | return loss 156 | 157 | def _generate_output( 158 | self, 159 | boxes: List[Tuple[int, int, int, int]], 160 | loss: Dict[str, torch.Tensor], 161 | top_k: int = 1 162 | ) -> List[Dict]: 163 | """ Generate top_k predictions as an output of the model. 164 | Argument: 165 | boxes (List[Tuple[int, int, int, int]]): bounding boxes in format [(xmin, ymin, xmax, ymax)] 166 | loss (Dict[str, torch.Tensor]): predicted CLIP loss in format {"prompt_label": (B, )}. 167 | top_k (int): top k best matches will be returned. 168 | Use top_k = -1 to return all boxes in ranked order. 169 | Returns: 170 | outputs (List[Dict]): predicts in format: 171 | { 172 | : { 173 | "src": , 174 | "ranking": [{"loss": loss, "idx": idx, "rect": [xmin, ymin, xmax, ymax]}] 175 | } 176 | } 177 | """ 178 | output = {} 179 | top_k = min(top_k, len(boxes)) if top_k > 0 else len(boxes) 180 | for key, val in loss.items(): 181 | output[key] = {} 182 | if key in self.get_prompts_list(): 183 | output[key]["src"] = self.get_prompt(key).src 184 | if not isinstance(output[key]["src"], str): 185 | output[key]["src"] = key 186 | ranking = [] 187 | vals, ids = val.sort() 188 | for i in range(top_k): 189 | ranking.append({ 190 | "loss": vals[i], 191 | "idx": ids[i], 192 | "rect": boxes[ids[i]] 193 | }) 194 | output[key]["ranking"] = ranking 195 | return output 196 | -------------------------------------------------------------------------------- /requirements.txt: -------------------------------------------------------------------------------- 1 | torch 2 | opencv-python 3 | pytorch_clip_guided_loss 4 | numpy 5 | -------------------------------------------------------------------------------- /resources/preds.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/bes-dev/pytorch_clip_bbox/1cc8ef2bdec8201153c4f5d3dc25b7f88b09bad4/resources/preds.jpg -------------------------------------------------------------------------------- /setup.cfg: -------------------------------------------------------------------------------- 1 | [metadata] 2 | description-file = README.md -------------------------------------------------------------------------------- /setup.py: -------------------------------------------------------------------------------- 1 | """ 2 | Copyright 2021 by Sergei Belousov 3 | 4 | Licensed under the Apache License, Version 2.0 (the "License"); 5 | you may not use this file except in compliance with the License. 6 | You may obtain a copy of the License at 7 | 8 | http://www.apache.org/licenses/LICENSE-2.0 9 | 10 | Unless required by applicable law or agreed to in writing, software 11 | distributed under the License is distributed on an "AS IS" BASIS, 12 | WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 13 | See the License for the specific language governing permissions and 14 | limitations under the License. 15 | """ 16 | from setuptools import setup, find_packages 17 | 18 | readme = open('README.md').read() 19 | 20 | VERSION = '2021.12.25.0' 21 | 22 | requirements = [ 23 | 'wheel', 24 | 'Cython', 25 | 'cython', 26 | 'torch', 27 | 'opencv-python', 28 | 'numpy', 29 | 'pytorch_clip_guided_loss' 30 | ] 31 | 32 | setup( 33 | # Metadata 34 | name='pytorch_clip_bbox', 35 | version=VERSION, 36 | author='Sergei Belousov aka BeS', 37 | author_email='sergei.o.belousov@gmail.com', 38 | description='Pytorch implementation of the CLIP guided bbox refinement for Object Detection.', 39 | long_description=readme, 40 | long_description_content_type='text/markdown', 41 | 42 | # Package info 43 | packages=find_packages(exclude=('*test*',)), 44 | 45 | # 46 | zip_safe=True, 47 | install_requires=requirements, 48 | 49 | # Classifiers 50 | classifiers=[ 51 | 'Programming Language :: Python :: 3', 52 | ], 53 | 54 | 55 | # install .json configs 56 | package_data={ 57 | # "pytorch_clip_guided_loss": ["configs/*.yml"] 58 | }, 59 | include_package_data=True 60 | ) 61 | --------------------------------------------------------------------------------