{"url":"/sota/unsupervised-video-object-segmentation-on-10","task":{"name":"Unsupervised Video Object Segmentation","url":"/task/unsupervised-video-object-segmentation","note":null},"dataset":{"name":"DAVIS 2016 val","url":"/dataset/davis-2016"},"category":"Computer Vision","categories":["Computer Vision"],"category_note":null,"description":"The unsupervised scenario assumes that the user does not interact with the algorithm to obtain the segmentation masks. Methods should provide a set of object candidates with no overlapping pixels that span through the whole video sequence. This set of objects should contain at least the objects that capture human attention when watching the whole video sequence i.e objects that are more likely to be followed by human gaze.","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":["G","J","F"],"metric_direction":{"note":"inferred from the metric name only (the archive records no direction); null = not inferred, chart draws points only","by_metric":{"G":null,"J":null,"F":null}},"counts":{"rows":25,"rows_with_code":17,"rows_with_paper_page":25,"rows_dated":24,"rows_using_additional_data":2},"rows":[{"rank_in_archive_order":1,"model":"GSANet","metrics":{"F":"89.6","G":"88.9","J":"88.3"},"uses_additional_data":false,"paper_date":"2023-03-15","paper":"/paper/guided-slot-attention-for-unsupervised-video","paper_url":"https://arxiv.org/abs/2303.08314v3","paper_title":"Guided Slot Attention for Unsupervised Video Object Segmentation","code":"https://github.com/hydragon516/gsanet","n_code_links":1,"syntology":null},{"rank_in_archive_order":2,"model":"DEVA (DIS)","metrics":{"F":"90.2","G":"88.9","J":"87.6"},"uses_additional_data":true,"paper_date":"2023-09-07","paper":"/paper/tracking-anything-with-decoupled-video","paper_url":"https://arxiv.org/abs/2309.03903v1","paper_title":"Tracking Anything with Decoupled Video Segmentation","code":"https://github.com/hkchengrex/Tracking-Anything-with-DEVA","n_code_links":1,"syntology":{"n_ran":7,"n_unverified":3,"n_samples":10,"n_pointer_only_licence":10}},{"rank_in_archive_order":3,"model":"FakeFlow","metrics":{"F":"89.0","G":"88.5","J":"88.0"},"uses_additional_data":false,"paper_date":"2024-07-16","paper":"/paper/improving-unsupervised-video-object-1","paper_url":"https://arxiv.org/abs/2407.11714v1","paper_title":"Improving Unsupervised Video Object Segmentation via Fake Flow Generation","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":4,"model":"DPA","metrics":{"F":"88.4","G":"87.6","J":"86.8"},"uses_additional_data":false,"paper_date":"2022-11-22","paper":"/paper/domain-alignment-and-temporal-aggregation-for","paper_url":"https://arxiv.org/abs/2211.12036v3","paper_title":"Dual Prototype Attention for Unsupervised Video Object Segmentation","code":"https://github.com/hydragon516/dpa","n_code_links":1,"syntology":null},{"rank_in_archive_order":5,"model":"AMP","metrics":{"F":"87.5","G":"87.3","J":"87.1"},"uses_additional_data":false,"paper_date":"2023-03-18","paper":"/paper/adaptive-multi-source-predictor-for-zero-shot","paper_url":"https://arxiv.org/abs/2303.10383v2","paper_title":"Adaptive Multi-source Predictor for Zero-shot Video Object Segmentation","code":"https://github.com/xiaoqi-zhao-dlut/multi-source-aps-zvos","n_code_links":1,"syntology":null},{"rank_in_archive_order":6,"model":"TMO (MiT-b1)","metrics":{"F":"87.8","G":"87.2","J":"86.6"},"uses_additional_data":false,"paper_date":"2022-09-04","paper":"/paper/treating-motion-as-option-to-reduce-motion","paper_url":"https://arxiv.org/abs/2209.03138v5","paper_title":"Treating Motion as Option to Reduce Motion Dependency in Unsupervised Video Object Segmentation","code":"https://github.com/suhwan-cho/tmo","n_code_links":2,"syntology":{"n_ran":0,"n_unverified":4,"n_samples":4,"n_pointer_only_licence":0}},{"rank_in_archive_order":7,"model":"TMO (RN-101)","metrics":{"F":"86.6","G":"86.1","J":"85.6"},"uses_additional_data":false,"paper_date":"2022-09-04","paper":"/paper/treating-motion-as-option-to-reduce-motion","paper_url":"https://arxiv.org/abs/2209.03138v5","paper_title":"Treating Motion as Option to Reduce Motion Dependency in Unsupervised Video Object Segmentation","code":"https://github.com/suhwan-cho/tmo","n_code_links":2,"syntology":{"n_ran":0,"n_unverified":4,"n_samples":4,"n_pointer_only_licence":0}},{"rank_in_archive_order":8,"model":"D2Conv3D","metrics":{"F":"86.5","G":"86.0","J":"85.5"},"uses_additional_data":false,"paper_date":"2021-11-15","paper":"/paper/d2conv3d-dynamic-dilated-convolutions-for","paper_url":"https://openaccess.thecvf.com/content/WACV2022/html/Schmidt_D2Conv3D_Dynamic_Dilated_Convolutions_for_Object_Segmentation_in_Videos_WACV_2022_paper.html","paper_title":"D2Conv3D: Dynamic Dilated Convolutions for Object Segmentation in Videos","code":"https://github.com/schmiddo/d2conv3d","n_code_links":2,"syntology":null},{"rank_in_archive_order":9,"model":"PMN","metrics":{"F":"86.4","G":"85.9","J":"85.4"},"uses_additional_data":false,"paper_date":"2022-09-08","paper":"/paper/unsupervised-video-object-segmentation-via","paper_url":"https://arxiv.org/abs/2209.03712v1","paper_title":"Unsupervised Video Object Segmentation via Prototype Memory Network","code":"https://github.com/Hydragon516/PMN","n_code_links":1,"syntology":null},{"rank_in_archive_order":10,"model":"DLDA","metrics":{"F":"86.6","G":"85.75","J":"84.9"},"uses_additional_data":true,"paper_date":"2024-01-25","paper":"/paper/self-supervised-video-object-segmentation-1","paper_url":"https://arxiv.org/abs/2401.13937v2","paper_title":"Self-supervised Video Object Segmentation with Distillation Learning of Deformable Attention","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":11,"model":"IMP","metrics":{"F":"86.7","G":"85.6","J":"84.5"},"uses_additional_data":false,"paper_date":"2021-12-23","paper":"/paper/iteratively-selecting-an-easy-reference-frame","paper_url":"https://arxiv.org/abs/2112.12402v1","paper_title":"Iteratively Selecting an Easy Reference Frame Makes Unsupervised Video Object Segmentation Easier","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":12,"model":"RTNet","metrics":{"F":"84.7","G":"85.2","J":"85.6"},"uses_additional_data":false,"paper_date":"2021-06-19","paper":"/paper/reciprocal-transformations-for-unsupervised","paper_url":"http://openaccess.thecvf.com//content/CVPR2021/html/Ren_Reciprocal_Transformations_for_Unsupervised_Video_Object_Segmentation_CVPR_2021_paper.html","paper_title":"Reciprocal Transformations for Unsupervised Video Object Segmentation","code":"https://github.com/OliverRensu/RTNet","n_code_links":1,"syntology":null},{"rank_in_archive_order":13,"model":"TransportNet","metrics":{"F":"85.0","G":"84.8","J":"84.5"},"uses_additional_data":false,"paper_date":"2021-01-01","paper":"/paper/deep-transport-network-for-unsupervised-video","paper_url":"http://openaccess.thecvf.com//content/ICCV2021/html/Zhang_Deep_Transport_Network_for_Unsupervised_Video_Object_Segmentation_ICCV_2021_paper.html","paper_title":"Deep Transport Network for Unsupervised Video Object Segmentation","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":14,"model":"AMC-Net","metrics":{"F":"84.6","G":"84.6","J":"84.5"},"uses_additional_data":false,"paper_date":"2021-01-01","paper":"/paper/learning-motion-appearance-co-attention-for","paper_url":"http://openaccess.thecvf.com//content/ICCV2021/html/Yang_Learning_Motion-Appearance_Co-Attention_for_Zero-Shot_Video_Object_Segmentation_ICCV_2021_paper.html","paper_title":"Learning Motion-Appearance Co-Attention for Zero-Shot Video Object Segmentation","code":"https://github.com/isyangshu/amc-net","n_code_links":1,"syntology":null},{"rank_in_archive_order":15,"model":"3DC-Seg","metrics":{"F":"84.7","G":"84.5","J":"84.3"},"uses_additional_data":false,"paper_date":"2020-08-26","paper":"/paper/making-a-case-for-3d-convolutions-for-object","paper_url":"https://arxiv.org/abs/2008.11516v2","paper_title":"Making a Case for 3D Convolutions for Object Segmentation in Videos","code":"https://github.com/sabarim/3DC-Seg","n_code_links":1,"syntology":{"n_ran":3,"n_unverified":6,"n_samples":9,"n_pointer_only_licence":0}},{"rank_in_archive_order":16,"model":"F2Net","metrics":{"F":"54.4","G":"83.7","J":"83.1"},"uses_additional_data":false,"paper_date":"2020-12-04","paper":"/paper/f2net-learning-to-focus-on-the-foreground-for","paper_url":"https://arxiv.org/abs/2012.02534v1","paper_title":"F2Net: Learning to Focus on the Foreground for Unsupervised Video Object Segmentation","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":17,"model":"FSNet","metrics":{"F":"83.1","G":"83.3","J":"83.4"},"uses_additional_data":false,"paper_date":"2021-08-06","paper":"/paper/full-duplex-strategy-for-video-object","paper_url":"https://arxiv.org/abs/2108.03151v3","paper_title":"Full-Duplex Strategy for Video Object Segmentation","code":"https://github.com/GewelsJI/FSNet","n_code_links":1,"syntology":{"n_ran":7,"n_unverified":1,"n_samples":8,"n_pointer_only_licence":0}},{"rank_in_archive_order":18,"model":"DFNet","metrics":{"F":"81.8","G":"82.6","J":"83.4"},"uses_additional_data":false,"paper_date":"2020-08-04","paper":"/paper/learning-discriminative-feature-with-crf-for","paper_url":"https://arxiv.org/abs/2008.01270v1","paper_title":"Learning Discriminative Feature with CRF for Unsupervised Video Object Segmentation","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":19,"model":"MATNet","metrics":{"F":"80.7","G":"81.6","J":"82.4"},"uses_additional_data":false,"paper_date":"2020-03-09","paper":"/paper/motion-attentive-transition-for-zero-shot","paper_url":"https://arxiv.org/abs/2003.04253v3","paper_title":"Motion-Attentive Transition for Zero-Shot Video Object Segmentation","code":"https://github.com/tfzhou/MATNet","n_code_links":1,"syntology":null},{"rank_in_archive_order":20,"model":"WCS-Net","metrics":{"F":"80.7","G":"81.5","J":"82.2"},"uses_additional_data":false,"paper_date":null,"paper":"/paper/unsupervised-video-object-segmentation-with-2","paper_url":"https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/2189_ECCV_2020_paper.php","paper_title":"Unsupervised Video Object Segmentation with Joint Hotspot Tracking","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":21,"model":"AD-Net","metrics":{"F":"80.5","G":"81.1","J":"81.7"},"uses_additional_data":false,"paper_date":"2019-10-24","paper":"/paper/anchor-diffusion-for-unsupervised-video-1","paper_url":"https://arxiv.org/abs/1910.10895v1","paper_title":"Anchor Diffusion for Unsupervised Video Object Segmentation","code":"https://github.com/yz93/anchor-diff-VOS","n_code_links":1,"syntology":null},{"rank_in_archive_order":22,"model":"COSNet","metrics":{"F":"79.4","G":"80.0","J":"80.5"},"uses_additional_data":false,"paper_date":"2020-01-19","paper":"/paper/see-more-know-more-unsupervised-video-object-1","paper_url":"https://arxiv.org/abs/2001.06810v1","paper_title":"See More, Know More: Unsupervised Video Object Segmentation with Co-Attention Siamese Networks","code":"https://github.com/carrierlxk/COSNet","n_code_links":1,"syntology":null},{"rank_in_archive_order":23,"model":"AGNN","metrics":{"F":"79.1","G":"79.9","J":"80.7"},"uses_additional_data":false,"paper_date":"2020-01-19","paper":"/paper/zero-shot-video-object-segmentation-via-1","paper_url":"https://arxiv.org/abs/2001.06807v1","paper_title":"Zero-Shot Video Object Segmentation via Attentive Graph Neural Networks","code":"https://github.com/carrierlxk/AGNN","n_code_links":1,"syntology":{"n_ran":4,"n_unverified":1,"n_samples":5,"n_pointer_only_licence":5}},{"rank_in_archive_order":24,"model":"AGS","metrics":{"F":"77.4","G":"78.6","J":"79.7"},"uses_additional_data":false,"paper_date":"2019-06-01","paper":"/paper/learning-unsupervised-video-object","paper_url":"http://openaccess.thecvf.com/content_CVPR_2019/html/Wang_Learning_Unsupervised_Video_Object_Segmentation_Through_Visual_Attention_CVPR_2019_paper.html","paper_title":"Learning Unsupervised Video Object Segmentation Through Visual Attention","code":"https://github.com/wenguanwang/AGS","n_code_links":1,"syntology":null},{"rank_in_archive_order":25,"model":"PDB","metrics":{"F":"74.5","G":"75.9","J":"77.2"},"uses_additional_data":false,"paper_date":"2018-09-01","paper":"/paper/pyramid-dilated-deeper-convlstm-for-video","paper_url":"http://openaccess.thecvf.com/content_ECCV_2018/html/Hongmei_Song_Pseudo_Pyramid_Deeper_ECCV_2018_paper.html","paper_title":"Pyramid Dilated Deeper ConvLSTM for Video Salient Object Detection","code":null,"n_code_links":0,"syntology":null}],"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":6,"rows_with_any_sample_ran":4,"distinct_papers_with_graph_line":5,"distinct_papers_with_any_sample_ran":4,"samples_over_distinct_papers":{"n_ran":21,"n_unverified":15,"n_samples":36,"n_pointer_only_licence":15,"note":"each paper (arXiv id) counted once, however many rows it is behind; this is the page-level figure"},"samples_row_weighted":{"n_ran":21,"n_unverified":19,"n_samples":40,"n_pointer_only_licence":15,"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"}}}