{"url":"/task/unsupervised-semantic-segmentation-with","name":"Unsupervised Semantic Segmentation with Language-image Pre-training","slug":"unsupervised-semantic-segmentation-with","description_markdown":"A segmentation task which does not utilise any human-level supervision for semantic segmentation except for a backbone which is initialised with features pre-trained with image-level labels.","categories":[{"name":"Computer Vision","url":"/area/computer-vision"}],"source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","slug_source":"archive_url"},"counts":{"papers_tagged":14,"papers_with_code":14,"benchmarks":12,"benchmark_tables_in_archive":12,"benchmark_tables_shown":12,"benchmark_tables_withheld_as_spam":0,"benchmark_definition":"a leaderboard table with at least one row; benchmark_tables_shown also counts the zero-row tables; benchmark_tables_in_archive adds the tables withheld as spam","datasets":7,"subtasks":0,"parent_tasks":1},"benchmarks":[{"leaderboard":"/sota/unsupervised-semantic-segmentation-with-4","slug":"unsupervised-semantic-segmentation-with-4","dataset":"ADE20K","dataset_url":"/dataset/ade20k","rows_in_archive":13,"metrics":["Mean IoU (val)"],"first_row_in_archive_order":{"model":"CorrCLIP","paper_title":"CorrCLIP: Reconstructing Correlations in CLIP with Off-the-Shelf Foundation Models for Open-Vocabulary Semantic Segmentation","paper_url":"/paper/corrclip-reconstructing-correlations-in-clip","paper_date":"2024-11-15","arxiv_id":"2411.10086","code_links":[{"title":"zdk258/CorrCLIP","url":"https://github.com/zdk258/CorrCLIP"}],"syntology":{"n":9,"n_ran":0,"n_unverified":9,"n_pointer_only":9}}},{"leaderboard":"/sota/unsupervised-semantic-segmentation-with-10","slug":"unsupervised-semantic-segmentation-with-10","dataset":"COCO-Object","dataset_url":null,"rows_in_archive":12,"metrics":["mIoU"],"first_row_in_archive_order":{"model":"CorrCLIP","paper_title":"CorrCLIP: Reconstructing Correlations in CLIP with Off-the-Shelf Foundation Models for Open-Vocabulary Semantic Segmentation","paper_url":"/paper/corrclip-reconstructing-correlations-in-clip","paper_date":"2024-11-15","arxiv_id":"2411.10086","code_links":[{"title":"zdk258/CorrCLIP","url":"https://github.com/zdk258/CorrCLIP"}],"syntology":{"n":9,"n_ran":0,"n_unverified":9,"n_pointer_only":9}}},{"leaderboard":"/sota/unsupervised-semantic-segmentation-with-3","slug":"unsupervised-semantic-segmentation-with-3","dataset":"Cityscapes val","dataset_url":"/dataset/cityscapes","rows_in_archive":12,"metrics":["mIoU","pixel accuracy"],"first_row_in_archive_order":{"model":"CorrCLIP","paper_title":"CorrCLIP: Reconstructing Correlations in CLIP with Off-the-Shelf Foundation Models for Open-Vocabulary Semantic Segmentation","paper_url":"/paper/corrclip-reconstructing-correlations-in-clip","paper_date":"2024-11-15","arxiv_id":"2411.10086","code_links":[{"title":"zdk258/CorrCLIP","url":"https://github.com/zdk258/CorrCLIP"}],"syntology":{"n":9,"n_ran":0,"n_unverified":9,"n_pointer_only":9}}},{"leaderboard":"/sota/unsupervised-semantic-segmentation-with-8","slug":"unsupervised-semantic-segmentation-with-8","dataset":"PASCAL Context-59","dataset_url":null,"rows_in_archive":12,"metrics":["mIoU"],"first_row_in_archive_order":{"model":"CorrCLIP","paper_title":"CorrCLIP: Reconstructing Correlations in CLIP with Off-the-Shelf Foundation Models for Open-Vocabulary Semantic Segmentation","paper_url":"/paper/corrclip-reconstructing-correlations-in-clip","paper_date":"2024-11-15","arxiv_id":"2411.10086","code_links":[{"title":"zdk258/CorrCLIP","url":"https://github.com/zdk258/CorrCLIP"}],"syntology":{"n":9,"n_ran":0,"n_unverified":9,"n_pointer_only":9}}},{"leaderboard":"/sota/unsupervised-semantic-segmentation-with-9","slug":"unsupervised-semantic-segmentation-with-9","dataset":"COCO-Stuff-171","dataset_url":"/dataset/coco-stuff","rows_in_archive":12,"metrics":["mIoU"],"first_row_in_archive_order":{"model":"CorrCLIP","paper_title":"CorrCLIP: Reconstructing Correlations in CLIP with Off-the-Shelf Foundation Models for Open-Vocabulary Semantic Segmentation","paper_url":"/paper/corrclip-reconstructing-correlations-in-clip","paper_date":"2024-11-15","arxiv_id":"2411.10086","code_links":[{"title":"zdk258/CorrCLIP","url":"https://github.com/zdk258/CorrCLIP"}],"syntology":{"n":9,"n_ran":0,"n_unverified":9,"n_pointer_only":9}}},{"leaderboard":"/sota/unsupervised-semantic-segmentation-with-11","slug":"unsupervised-semantic-segmentation-with-11","dataset":"PASCAL VOC","dataset_url":"/dataset/pascal-voc","rows_in_archive":10,"metrics":["mIoU"],"first_row_in_archive_order":{"model":"CorrCLIP","paper_title":"CorrCLIP: Reconstructing Correlations in CLIP with Off-the-Shelf Foundation Models for Open-Vocabulary Semantic Segmentation","paper_url":"/paper/corrclip-reconstructing-correlations-in-clip","paper_date":"2024-11-15","arxiv_id":"2411.10086","code_links":[{"title":"zdk258/CorrCLIP","url":"https://github.com/zdk258/CorrCLIP"}],"syntology":{"n":9,"n_ran":0,"n_unverified":9,"n_pointer_only":9}}},{"leaderboard":"/sota/unsupervised-semantic-segmentation-with-7","slug":"unsupervised-semantic-segmentation-with-7","dataset":"PascalVOC-20","dataset_url":"/dataset/pascal-voc","rows_in_archive":10,"metrics":["mIoU"],"first_row_in_archive_order":{"model":"CorrCLIP","paper_title":"CorrCLIP: Reconstructing Correlations in CLIP with Off-the-Shelf Foundation Models for Open-Vocabulary Semantic Segmentation","paper_url":"/paper/corrclip-reconstructing-correlations-in-clip","paper_date":"2024-11-15","arxiv_id":"2411.10086","code_links":[{"title":"zdk258/CorrCLIP","url":"https://github.com/zdk258/CorrCLIP"}],"syntology":{"n":9,"n_ran":0,"n_unverified":9,"n_pointer_only":9}}},{"leaderboard":"/sota/unsupervised-semantic-segmentation-with-1","slug":"unsupervised-semantic-segmentation-with-1","dataset":"COCO-Stuff-27","dataset_url":"/dataset/coco-stuff","rows_in_archive":4,"metrics":["mIoU","pixel accuracy"],"first_row_in_archive_order":{"model":"ReCo+","paper_title":"ReCo: Retrieve and Co-segment for Zero-shot Transfer","paper_url":"/paper/reco-retrieve-and-co-segment-for-zero-shot-1","paper_date":"2022-06-14","arxiv_id":"2206.07045","code_links":[{"title":"NoelShin/reco","url":"https://github.com/NoelShin/reco"},{"title":"noelshin/namedmask","url":"https://github.com/noelshin/namedmask"}],"syntology":null}},{"leaderboard":"/sota/unsupervised-semantic-segmentation-with-12","slug":"unsupervised-semantic-segmentation-with-12","dataset":"PASCAL Context-60","dataset_url":null,"rows_in_archive":4,"metrics":["mIoU"],"first_row_in_archive_order":{"model":"CorrCLIP","paper_title":"CorrCLIP: Reconstructing Correlations in CLIP with Off-the-Shelf Foundation Models for Open-Vocabulary Semantic Segmentation","paper_url":"/paper/corrclip-reconstructing-correlations-in-clip","paper_date":"2024-11-15","arxiv_id":"2411.10086","code_links":[{"title":"zdk258/CorrCLIP","url":"https://github.com/zdk258/CorrCLIP"}],"syntology":{"n":9,"n_ran":0,"n_unverified":9,"n_pointer_only":9}}},{"leaderboard":"/sota/unsupervised-semantic-segmentation-with-2","slug":"unsupervised-semantic-segmentation-with-2","dataset":"KITTI-STEP","dataset_url":"/dataset/kitti-step","rows_in_archive":3,"metrics":["mIoU","pixel accuracy"],"first_row_in_archive_order":{"model":"ReCo+","paper_title":"ReCo: Retrieve and Co-segment for Zero-shot Transfer","paper_url":"/paper/reco-retrieve-and-co-segment-for-zero-shot-1","paper_date":"2022-06-14","arxiv_id":"2206.07045","code_links":[{"title":"NoelShin/reco","url":"https://github.com/NoelShin/reco"},{"title":"noelshin/namedmask","url":"https://github.com/noelshin/namedmask"}],"syntology":null}},{"leaderboard":"/sota/unsupervised-semantic-segmentation-with-5","slug":"unsupervised-semantic-segmentation-with-5","dataset":"COCO (Common Objects in Context)","dataset_url":"/dataset/coco","rows_in_archive":1,"metrics":["Mean IoU (val)"],"first_row_in_archive_order":{"model":"CLIPpy ViT-B","paper_title":"Perceptual Grouping in Contrastive Vision-Language Models","paper_url":"/paper/perceptual-grouping-in-vision-language-models","paper_date":"2022-10-18","arxiv_id":"2210.09996","code_links":[{"title":"kahnchana/clippy","url":"https://github.com/kahnchana/clippy"},{"title":"jongwoopark7978/LVNet","url":"https://github.com/jongwoopark7978/LVNet"}],"syntology":null}},{"leaderboard":"/sota/unsupervised-semantic-segmentation-with-6","slug":"unsupervised-semantic-segmentation-with-6","dataset":"PASCAL VOC 2007","dataset_url":"/dataset/pascal-voc-2007","rows_in_archive":1,"metrics":["Mean IoU (val)"],"first_row_in_archive_order":{"model":"CLIPpy ViT-B","paper_title":"Perceptual Grouping in Contrastive Vision-Language Models","paper_url":"/paper/perceptual-grouping-in-vision-language-models","paper_date":"2022-10-18","arxiv_id":"2210.09996","code_links":[{"title":"kahnchana/clippy","url":"https://github.com/kahnchana/clippy"},{"title":"jongwoopark7978/LVNet","url":"https://github.com/jongwoopark7978/LVNet"}],"syntology":null}}],"datasets":[{"url":"/dataset/coco","name":"COCO (Common Objects in Context)","full_name":"Common Objects in Context","num_papers_in_archive":11922},{"url":"/dataset/cityscapes","name":"Cityscapes","full_name":"","num_papers_in_archive":3702},{"url":"/dataset/ade20k","name":"ADE20K","full_name":"","num_papers_in_archive":1213},{"url":"/dataset/coco-stuff","name":"COCO-Stuff","full_name":"Common Objects in COntext-stuff","num_papers_in_archive":338},{"url":"/dataset/pascal-voc","name":"PASCAL VOC","full_name":"PASCAL Visual Object Classes Challenge","num_papers_in_archive":198},{"url":"/dataset/pascal-voc-2007","name":"PASCAL VOC 2007","full_name":"PASCAL VOC 2007","num_papers_in_archive":126},{"url":"/dataset/kitti-step","name":"KITTI-STEP","full_name":"","num_papers_in_archive":24}],"subtasks":[],"parent_tasks":[{"url":"/task/unsupervised-semantic-segmentation","name":"Unsupervised Semantic Segmentation"}],"papers":{"order":"repositories listed in the archive (desc), then date (desc); the archive holds no stars","population":"papers tagged with this task that list at least one repository in the archive","shown":14,"of":14,"tagged_in_all":14,"items":[{"url":"/paper/groupvit-semantic-segmentation-emerges-from","title":"GroupViT: Semantic Segmentation Emerges from Text Supervision","date":"2022-02-22","arxiv_id":"2202.11094","repositories_listed":6,"syntology":null},{"url":"/paper/perceptual-grouping-in-vision-language-models","title":"Perceptual Grouping in Contrastive Vision-Language Models","date":"2022-10-18","arxiv_id":"2210.09996","repositories_listed":2,"syntology":null},{"url":"/paper/reco-retrieve-and-co-segment-for-zero-shot-1","title":"ReCo: Retrieve and Co-segment for Zero-shot Transfer","date":"2022-06-14","arxiv_id":"2206.07045","repositories_listed":2,"syntology":null},{"url":"/paper/textregion-text-aligned-region-tokens-from","title":"TextRegion: Text-Aligned Region Tokens from Frozen Image-Text Models","date":"2025-05-29","arxiv_id":"2505.23769","repositories_listed":1,"syntology":null},{"url":"/paper/cosmos-cross-modality-self-distillation-for","title":"COSMOS: Cross-Modality Self-Distillation for Vision Language Pre-training","date":"2024-12-02","arxiv_id":"2412.01814","repositories_listed":1,"syntology":null},{"url":"/paper/itaclip-boosting-training-free-semantic","title":"ITACLIP: Boosting Training-Free Semantic Segmentation with Image, Text, and Architectural Enhancements","date":"2024-11-18","arxiv_id":"2411.12044","repositories_listed":1,"syntology":null},{"url":"/paper/corrclip-reconstructing-correlations-in-clip","title":"CorrCLIP: Reconstructing Correlations in CLIP with Off-the-Shelf Foundation Models for Open-Vocabulary Semantic Segmentation","date":"2024-11-15","arxiv_id":"2411.10086","repositories_listed":1,"syntology":{"n":9,"n_ran":0,"n_unverified":9,"n_pointer_only":9}},{"url":"/paper/harnessing-vision-foundation-models-for-high","title":"Harnessing Vision Foundation Models for High-Performance, Training-Free Open Vocabulary Segmentation","date":"2024-11-14","arxiv_id":"2411.09219","repositories_listed":1,"syntology":null},{"url":"/paper/proxyclip-proxy-attention-improves-clip-for","title":"ProxyCLIP: Proxy Attention Improves CLIP for Open-Vocabulary Segmentation","date":"2024-08-09","arxiv_id":"2408.04883","repositories_listed":1,"syntology":{"n":9,"n_ran":3,"n_unverified":6,"n_pointer_only":9}},{"url":"/paper/ttd-text-tag-self-distillation-enhancing","title":"TTD: Text-Tag Self-Distillation Enhancing Image-Text Alignment in CLIP to Alleviate Single Tag Bias","date":"2024-03-30","arxiv_id":"2404.00384","repositories_listed":1,"syntology":{"n":1,"n_ran":1,"n_unverified":0,"n_pointer_only":1}},{"url":"/paper/tagalign-improving-vision-language-alignment","title":"TagAlign: Improving Vision-Language Alignment with Multi-Tag Classification","date":"2023-12-21","arxiv_id":"2312.14149","repositories_listed":1,"syntology":null},{"url":"/paper/tagclip-a-local-to-global-framework-to","title":"TagCLIP: A Local-to-Global Framework to Enhance Open-Vocabulary Multi-Label Classification of CLIP Without Training","date":"2023-12-20","arxiv_id":"2312.12828","repositories_listed":1,"syntology":null},{"url":"/paper/learning-to-generate-text-grounded-mask-for","title":"Learning to Generate Text-grounded Mask for Open-world Semantic Segmentation from Only Image-Text Pairs","date":"2022-12-01","arxiv_id":"2212.00785","repositories_listed":1,"syntology":{"n":8,"n_ran":1,"n_unverified":7,"n_pointer_only":0}},{"url":"/paper/denseclip-extract-free-dense-labels-from-clip","title":"Extract Free Dense Labels from CLIP","date":"2021-12-02","arxiv_id":"2112.01071","repositories_listed":1,"syntology":null}],"syntology_records":4,"syntology_note":"a paper without a record is not a recorded non-run: it may lack an arXiv id or simply be absent from the graph layer"},"description_links":{"kept":0,"unwrapped_to_text":0,"bare_urls_linked":0,"relative_images_dropped":0,"rule":"internal links are kept only when the target slug exists in the catalog"},"syntology":{"read_at":"2026-09-24T18:15:14+00:00","claim":"Per-sample execution status on synthesized fixtures ('ran N of M samples'); not a correctness claim and not a ranking signal.","status_vocabulary":{"ran_honours":"ran, honoured the contract we drafted","ran_violates":"ran, violated the contract we drafted","ran_draft_wrong":"ran; our contract draft was wrong, not the code","ran_fixture":"ran; our fixture could not drive it","ran":"ran on a synthesized input","unverified":"unverified (harvested, no recorded run)"}},"not_shown":{"libraries":"the archive has no per-task library table","trend_sparklines":"the Trend column of the benchmarks table was a rendered image; it is not in the archive","social_and_latest_sorts":"stars and social signals are not in the archive"}}