{"url":"/sota/semantic-correspondence-on-spair-71k","task":{"name":"Semantic correspondence","url":"/task/semantic-correspondence","note":null},"dataset":{"name":"SPair-71k","url":"/dataset/spair-71k"},"category":"Computer Vision","categories":["Computer Vision"],"category_note":null,"description":"The task of semantic correspondence aims to establish reliable visual correspondence between different instances of the same object category.","description_from":"task","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","rank":"the archive's row order at snapshot; not re-ranked","rows_end_at":"2025-07-28","rows_withheld_as_spam":0,"metric_values":"the archive's strings, untouched"},"metrics":["PCK"],"metric_direction":{"note":"inferred from the metric name only (the archive records no direction); null = not inferred, chart draws points only","by_metric":{"PCK":null}},"counts":{"rows":22,"rows_with_code":21,"rows_with_paper_page":22,"rows_dated":22,"rows_using_additional_data":0},"rows":[{"rank_in_archive_order":1,"model":"GeoAware-SC (Supervised, AP-10K P.T.)","metrics":{"PCK":"85.6"},"uses_additional_data":false,"paper_date":"2023-11-28","paper":"/paper/telling-left-from-right-identifying-geometry","paper_url":"https://arxiv.org/abs/2311.17034v2","paper_title":"Telling Left from Right: Identifying Geometry-Aware Semantic Correspondence","code":"https://github.com/Junyi42/geoaware-sc","n_code_links":1,"syntology":{"n_ran":11,"n_unverified":4,"n_samples":15,"n_pointer_only_licence":15}},{"rank_in_archive_order":2,"model":"DINOv2","metrics":{"PCK":"85.2"},"uses_additional_data":false,"paper_date":"2025-05-23","paper":"/paper/semantic-correspondence-unified-benchmarking","paper_url":"https://arxiv.org/abs/2505.18060v1","paper_title":"Semantic Correspondence: Unified Benchmarking and a Strong Baseline","code":"https://github.com/Visual-AI/Semantic-Correspondence","n_code_links":1,"syntology":null},{"rank_in_archive_order":3,"model":"GeoAware-SC (Supervised)","metrics":{"PCK":"82.9"},"uses_additional_data":false,"paper_date":"2023-11-28","paper":"/paper/telling-left-from-right-identifying-geometry","paper_url":"https://arxiv.org/abs/2311.17034v2","paper_title":"Telling Left from Right: Identifying Geometry-Aware Semantic Correspondence","code":"https://github.com/Junyi42/geoaware-sc","n_code_links":1,"syntology":{"n_ran":11,"n_unverified":4,"n_samples":15,"n_pointer_only_licence":15}},{"rank_in_archive_order":4,"model":"SD+DINO (Supervised)","metrics":{"PCK":"74.6"},"uses_additional_data":false,"paper_date":"2023-05-24","paper":"/paper/a-tale-of-two-features-stable-diffusion","paper_url":"https://arxiv.org/abs/2305.15347v2","paper_title":"A Tale of Two Features: Stable Diffusion Complements DINO for Zero-Shot Semantic Correspondence","code":"https://github.com/Junyi42/sd-dino","n_code_links":1,"syntology":null},{"rank_in_archive_order":5,"model":"GeoAware-SC + CleanDIFT (Zero-Shot)","metrics":{"PCK":"70.0"},"uses_additional_data":false,"paper_date":"2024-12-04","paper":"/paper/cleandift-diffusion-features-without-noise","paper_url":"https://arxiv.org/abs/2412.03439v1","paper_title":"CleanDIFT: Diffusion Features without Noise","code":"https://github.com/CompVis/cleandift","n_code_links":1,"syntology":{"n_ran":0,"n_unverified":5,"n_samples":5,"n_pointer_only_licence":5}},{"rank_in_archive_order":6,"model":"GeoAware-SC (Zero-Shot)","metrics":{"PCK":"68.5"},"uses_additional_data":false,"paper_date":"2023-11-28","paper":"/paper/telling-left-from-right-identifying-geometry","paper_url":"https://arxiv.org/abs/2311.17034v2","paper_title":"Telling Left from Right: Identifying Geometry-Aware Semantic Correspondence","code":"https://github.com/Junyi42/geoaware-sc","n_code_links":1,"syntology":{"n_ran":11,"n_unverified":4,"n_samples":15,"n_pointer_only_licence":15}},{"rank_in_archive_order":7,"model":"SD+DINO + CleanDIFT (Zero-Shot)","metrics":{"PCK":"64.8"},"uses_additional_data":false,"paper_date":"2024-12-04","paper":"/paper/cleandift-diffusion-features-without-noise","paper_url":"https://arxiv.org/abs/2412.03439v1","paper_title":"CleanDIFT: Diffusion Features without Noise","code":"https://github.com/CompVis/cleandift","n_code_links":1,"syntology":{"n_ran":0,"n_unverified":5,"n_samples":5,"n_pointer_only_licence":5}},{"rank_in_archive_order":8,"model":"IFCAT","metrics":{"PCK":"64.4"},"uses_additional_data":false,"paper_date":"2022-09-19","paper":"/paper/integrative-feature-and-cost-aggregation-with","paper_url":"https://arxiv.org/abs/2209.08742v2","paper_title":"Integrative Feature and Cost Aggregation with Transformers for Dense Correspondence","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":9,"model":"SD+DINO (Zero-shot)","metrics":{"PCK":"64.0"},"uses_additional_data":false,"paper_date":"2023-05-24","paper":"/paper/a-tale-of-two-features-stable-diffusion","paper_url":"https://arxiv.org/abs/2305.15347v2","paper_title":"A Tale of Two Features: Stable Diffusion Complements DINO for Zero-Shot Semantic Correspondence","code":"https://github.com/Junyi42/sd-dino","n_code_links":1,"syntology":null},{"rank_in_archive_order":10,"model":"DIFT + CleanDIFT (Zero-Shot)","metrics":{"PCK":"61.4"},"uses_additional_data":false,"paper_date":"2024-12-04","paper":"/paper/cleandift-diffusion-features-without-noise","paper_url":"https://arxiv.org/abs/2412.03439v1","paper_title":"CleanDIFT: Diffusion Features without Noise","code":"https://github.com/CompVis/cleandift","n_code_links":1,"syntology":{"n_ran":0,"n_unverified":5,"n_samples":5,"n_pointer_only_licence":5}},{"rank_in_archive_order":11,"model":"CATs++","metrics":{"PCK":"59.8"},"uses_additional_data":false,"paper_date":"2022-02-14","paper":"/paper/cats-boosting-cost-aggregation-with","paper_url":"https://arxiv.org/abs/2202.06817v2","paper_title":"CATs++: Boosting Cost Aggregation with Convolutions and Transformers","code":"https://github.com/KU-CVLAB/CATs-PlusPlus","n_code_links":1,"syntology":null},{"rank_in_archive_order":12,"model":"VAT (ECCV)","metrics":{"PCK":"55.5"},"uses_additional_data":false,"paper_date":"2022-07-22","paper":"/paper/cost-aggregation-with-4d-convolutional-swin","paper_url":"https://arxiv.org/abs/2207.10866v1","paper_title":"Cost Aggregation with 4D Convolutional Swin Transformer for Few-Shot Segmentation","code":"https://github.com/Seokju-Cho/Volumetric-Aggregation-Transformer","n_code_links":1,"syntology":{"n_ran":11,"n_unverified":9,"n_samples":20,"n_pointer_only_licence":2}},{"rank_in_archive_order":13,"model":"VAT","metrics":{"PCK":"54.2"},"uses_additional_data":false,"paper_date":"2021-12-22","paper":"/paper/cost-aggregation-is-all-you-need-for-few-shot","paper_url":"https://arxiv.org/abs/2112.11685v1","paper_title":"Cost Aggregation Is All You Need for Few-Shot Segmentation","code":"https://github.com/Seokju-Cho/Volumetric-Aggregation-Transformer","n_code_links":2,"syntology":null},{"rank_in_archive_order":14,"model":"TransforMatcher","metrics":{"PCK":"53.7"},"uses_additional_data":false,"paper_date":"2022-05-23","paper":"/paper/transformatcher-match-to-match-attention-for","paper_url":"https://arxiv.org/abs/2205.11634v1","paper_title":"TransforMatcher: Match-to-Match Attention for Semantic Correspondence","code":"https://github.com/wookiekim/transformatcher","n_code_links":1,"syntology":null},{"rank_in_archive_order":15,"model":"MMNet","metrics":{"PCK":"50.4"},"uses_additional_data":false,"paper_date":"2021-07-31","paper":"/paper/multi-scale-matching-networks-for-semantic","paper_url":"https://arxiv.org/abs/2108.00211v2","paper_title":"Multi-scale Matching Networks for Semantic Correspondence","code":"https://github.com/wintersun661/MMNet","n_code_links":1,"syntology":null},{"rank_in_archive_order":16,"model":"CATs","metrics":{"PCK":"49.9"},"uses_additional_data":false,"paper_date":"2021-06-04","paper":"/paper/semantic-correspondence-with-transformers","paper_url":"https://arxiv.org/abs/2106.02520v3","paper_title":"CATs: Cost Aggregation Transformers for Visual Correspondence","code":"https://github.com/SunghwanHong/Cost-Aggregation-transformers","n_code_links":1,"syntology":null},{"rank_in_archive_order":17,"model":"CHM","metrics":{"PCK":"46.3"},"uses_additional_data":false,"paper_date":"2021-03-31","paper":"/paper/convolutional-hough-matching-networks","paper_url":"https://arxiv.org/abs/2103.16831v1","paper_title":"Convolutional Hough Matching Networks","code":"https://github.com/juhongm999/chm","n_code_links":1,"syntology":null},{"rank_in_archive_order":18,"model":"LDMCorrespondences","metrics":{"PCK":"45.4"},"uses_additional_data":false,"paper_date":"2023-05-24","paper":"/paper/unsupervised-semantic-correspondence-using","paper_url":"https://arxiv.org/abs/2305.15581v2","paper_title":"Unsupervised Semantic Correspondence Using Stable Diffusion","code":"https://github.com/ubc-vision/LDM_correspondences","n_code_links":1,"syntology":{"n_ran":4,"n_unverified":8,"n_samples":12,"n_pointer_only_licence":0}},{"rank_in_archive_order":19,"model":"DHPF","metrics":{"PCK":"37.3"},"uses_additional_data":false,"paper_date":"2020-07-21","paper":"/paper/learning-to-compose-hypercolumns-for-visual","paper_url":"https://arxiv.org/abs/2007.10587v1","paper_title":"Learning to Compose Hypercolumns for Visual Correspondence","code":"https://github.com/juhongm999/dhpf","n_code_links":1,"syntology":{"n_ran":0,"n_unverified":5,"n_samples":5,"n_pointer_only_licence":0}},{"rank_in_archive_order":20,"model":"SCOT","metrics":{"PCK":"35.6"},"uses_additional_data":false,"paper_date":"2020-06-01","paper":"/paper/semantic-correspondence-as-an-optimal","paper_url":"http://openaccess.thecvf.com/content_CVPR_2020/html/Liu_Semantic_Correspondence_as_an_Optimal_Transport_Problem_CVPR_2020_paper.html","paper_title":"Semantic Correspondence as an Optimal Transport Problem","code":"https://github.com/csyanbin/SCOT","n_code_links":1,"syntology":null},{"rank_in_archive_order":21,"model":"ANCNet","metrics":{"PCK":"30.1"},"uses_additional_data":false,"paper_date":"2020-03-26","paper":"/paper/correspondence-networks-with-adaptive","paper_url":"https://arxiv.org/abs/2003.12059v1","paper_title":"Correspondence Networks with Adaptive Neighbourhood Consensus","code":"https://github.com/ActiveVisionLab/ANCNet","n_code_links":1,"syntology":{"n_ran":2,"n_unverified":9,"n_samples":11,"n_pointer_only_licence":0}},{"rank_in_archive_order":22,"model":"HPF","metrics":{"PCK":"28.2"},"uses_additional_data":false,"paper_date":"2019-08-18","paper":"/paper/hyperpixel-flow-semantic-correspondence-with","paper_url":"https://arxiv.org/abs/1908.06537v1","paper_title":"Hyperpixel Flow: Semantic Correspondence with Multi-layer Neural Features","code":"https://github.com/juhongm999/hpf","n_code_links":1,"syntology":{"n_ran":0,"n_unverified":14,"n_samples":14,"n_pointer_only_licence":0}}],"since_archive":{"claim":"Results that newer papers report for their own method, placed here by Syntology. A model pointed at the cell in the paper's own table; the number was read from that cell and checked against this leaderboard's metric, dataset, split and scale; an independent check that saw this leaderboard's other rows and every other leaderboard on the same dataset accepted it. Not reviewed by the paper's authors or by the archive's editors, and not ranked against the archive rows.","extraction_file_present":true,"measurement":{"test_papers":883,"papers_with_output":881,"judged_true":108,"judged":110,"wilson95_lower":0.9361,"measured_on":"2026-09-24","frozen_commit":"0e3de0df94"},"measurement_note":"blind adjudication of accepted entries on a held-out split of archive papers, rules frozen before the test","coverage":{"sentence":"Syntology has checked 6,795 of the 9,581 papers on this site that are newer than the archive; results from the others appear after they are checked.","complete":false,"papers_newer_than_archive":9581,"papers_checked":6795,"papers_extracted_not_yet_verified":0,"boards_without_verdict":2,"papers_not_yet_extracted":2785},"order":"newest first by month (arXiv date, else the arXiv-id month), then arXiv id descending","columns":[],"entries":[]},"syntology":{"read_at":"2026-09-24T18:15:14+00:00","claim":"Per row: N of M harvested code samples from that row's paper executed on a synthesized fixture; the other M-N are unverified. Not a reproduction of the row's number; not a correctness claim. n_pointer_only_licence counts samples the site points at rather than redistributes (a licence axis, independent of ran/unverified).","rows_with_graph_line":11,"rows_with_any_sample_ran":6,"distinct_papers_with_graph_line":7,"distinct_papers_with_any_sample_ran":4,"samples_over_distinct_papers":{"n_ran":28,"n_unverified":54,"n_samples":82,"n_pointer_only_licence":22,"note":"each paper (arXiv id) counted once, however many rows it is behind; this is the page-level figure"},"samples_row_weighted":{"n_ran":50,"n_unverified":72,"n_samples":122,"n_pointer_only_licence":62,"note":"row-weighted: a paper behind several rows is counted once per row; inflated relative to samples_over_distinct_papers by design, kept for readers summing the per-row syntology blocks"}}}