{"id":13488584,"url":"https://github.com/PaulCouairon/DiffCut","last_synced_at":"2025-03-28T01:36:34.796Z","repository":{"id":242783503,"uuid":"804073848","full_name":"PaulCouairon/DiffCut","owner":"PaulCouairon","description":"Official code for \"DiffCut: Catalyzing Zero-Shot Semantic Segmentation with Diffusion Features and Recursive Normalized Cut\", NeurIPS 2024","archived":false,"fork":false,"pushed_at":"2024-10-15T20:32:25.000Z","size":18783,"stargazers_count":20,"open_issues_count":1,"forks_count":2,"subscribers_count":1,"default_branch":"main","last_synced_at":"2024-10-17T07:04:49.232Z","etag":null,"topics":["computer-vision","image-segmentation","normalized-cuts","stable-diffusion","unsupervised-segmentation","zero-shot-segmentation"],"latest_commit_sha":null,"homepage":"https://diffcut-segmentation.github.io","language":"Jupyter Notebook","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"mit","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/PaulCouairon.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null}},"created_at":"2024-05-21T22:53:47.000Z","updated_at":"2024-10-15T20:32:28.000Z","dependencies_parsed_at":"2024-06-10T11:44:20.345Z","dependency_job_id":null,"html_url":"https://github.com/PaulCouairon/DiffCut","commit_stats":null,"previous_names":["paulcouairon/diffcut"],"tags_count":0,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/PaulCouairon%2FDiffCut","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/PaulCouairon%2FDiffCut/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/PaulCouairon%2FDiffCut/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/PaulCouairon%2FDiffCut/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/PaulCouairon","download_url":"https://codeload.github.com/PaulCouairon/DiffCut/tar.gz/refs/heads/main","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":222333976,"owners_count":16968058,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["computer-vision","image-segmentation","normalized-cuts","stable-diffusion","unsupervised-segmentation","zero-shot-segmentation"],"created_at":"2024-07-31T18:01:18.427Z","updated_at":"2025-03-28T01:36:34.789Z","avatar_url":"https://github.com/PaulCouairon.png","language":"Jupyter Notebook","funding_links":[],"categories":["Segmentation Detection Tracking"],"sub_categories":[],"readme":"\u003cdiv align=\"center\"\u003e\n\u003ch1\u003e\nDiffCut: Catalyzing Zero-Shot Semantic Segmentation with Diffusion Features and Recursive Normalized Cut\n\u003cbr\u003e\n\u003c/h1\u003e\n\n\u003ch2\u003e\nNeurIPS 2024\n\u003cbr\u003e\n\u003cbr\u003e\n\u003ca href=\"https://scholar.google.fr/citations?user=yQRnP7YAAAAJ\u0026hl=fr\"\u003ePaul Couairon\u003c/a\u003e\u0026ensp;\n\u003ca href=\"https://scholar.google.com/citations?user=lhp9mRgAAAAJ\u0026hl=en\"\u003eMustafa Shukor\u003c/a\u003e\u0026ensp;\n\n\u003ca href=\"https://fr.linkedin.com/in/jean-emmanuel-haugeard\"\u003eJean-Emmanuel Haugeard\u003c/a\u003e\u0026ensp;\n\u003ca href=\"https://cord.isir.upmc.fr\"\u003eMatthieu Cord\u003c/a\u003e\u0026ensp;\n\u003ca href=\"https://thome.isir.upmc.fr\"\u003eNicolas Thome\u003c/a\u003e\u0026ensp;\n\u003c/h2\u003e\n\n\n\u003cp\u003e\u003c/p\u003e\n\u003ca href=\"https://arxiv.org/abs/2406.02842v2\"\u003e\u003cimg\nsrc=\"https://img.shields.io/badge/arXiv-DiffCut-b31b1b.svg\" height=25em\u003e\u003c/a\u003e\n\u003ca href=\"https://diffcut-segmentation.github.io\"\u003e\u003cimg \nsrc=\"https://img.shields.io/static/v1?label=Project\u0026message=Website\u0026color=green\" height=25em\u003e\u003c/a\u003e\n\n\n![main_figure.png](./assets/main_figure.png)\n\n\u003c/div\u003e\n\n## Environment\n```\nconda create -n diffcut python=3.10\nconda activate diffcut\npip install -r requirements.txt\n```\n\nFor evaluation, install [detectron2](https://detectron2.readthedocs.io/en/latest/tutorials/install.html)\n```\npython -m pip install 'git+https://github.com/facebookresearch/detectron2.git'\n```\n\n## Demo\nTry our __DiffCut__ method by running the notebook ``diffcut.ipynb``\n\nVisualize the __semantic coherence__ of vision encoders (SD, CLIP, DINO...) with ``semantic_coherence.ipynb``\n\n\n\n\n## Evaluation\n\n### Datasets Preparation\nIn the paper, we evaluate DiffCut on 6 benchmarks: PASCAL VOC (20 classes + background), PASCAL Context (59 classes + background), COCO-Object (80 classes + background), COCO-Stuff (27 classes), Cityscapes (27 classes) and ADE20k (150 classes). See [Preparing Datasets for DiffCut](datasets/README.md).\n\n### Run Evaluation\n```\npython eval_diffcut.py --dataset_name Cityscapes --tau 0.5 --alpha 10 --refinement\npython eval_diffcut_openvoc.py --dataset_name VOC20 --tau 0.5 --alpha 10 --refinement\n```\n\n\n## Citation\n```\n@inproceedings{\n    couairon2024diffcut,\n    title={DiffCut: Catalyzing Zero-Shot Semantic Segmentation with Diffusion Features and Recursive Normalized Cut},\n    author={Paul Couairon and Mustafa Shukor and Jean-Emmanuel HAUGEARD and Matthieu Cord and Nicolas THOME},\n    booktitle={The Thirty-eighth Annual Conference on Neural Information Processing Systems},\n    year={2024},\n    url={https://openreview.net/forum?id=N0xNf9Qqmc}\n}\n```\n\n## Acknowledgements\nThis repo relies on the following projects:\n\n[Diffuse, Attend, and Segment: Unsupervised Zero-Shot Segmentation using Stable Diffusion](https://github.com/google/diffseg)\n\n[Emergent Correspondence from Image Diffusion](https://github.com/Tsingularity/dift)\n\n[Open Vocabulary Semantic Segmentation with Patch Aligned Contrastive Learning](https://arxiv.org/abs/2212.04994)\n\n[Convolutions Die Hard: Open-Vocabulary Segmentation with Single Frozen Convolutional CLIP](https://github.com/bytedance/fc-clip)\n\n[Cut and Learn for Unsupervised Image \u0026 Video Object Detection and Instance Segmentation](https://github.com/facebookresearch/CutLER)\n\n\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2FPaulCouairon%2FDiffCut","html_url":"https://awesome.ecosyste.ms/projects/github.com%2FPaulCouairon%2FDiffCut","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2FPaulCouairon%2FDiffCut/lists"}