{"url":"/sota/referring-expression-segmentation-on-refcocog","task":{"name":"Referring Expression Segmentation","url":"/task/referring-expression-segmentation","note":null},"dataset":{"name":"RefCOCOg-val","url":"/dataset/google-refexp"},"category":"Computer Vision","categories":["Computer Vision"],"category_note":null,"description":"The task aims at labeling the pixels of an image or video that represent an object instance referred by a linguistic expression. In particular, the referring expression (RE) must allow the identification of an individual object in a discourse or scene (the referent). REs unambiguously identify the target instance.","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":["Overall IoU","Mean IoU","IoU","mIoU"],"metric_direction":{"note":"inferred from the metric name only (the archive records no direction); null = not inferred, chart draws points only","by_metric":{"Overall IoU":"higher","Mean IoU":"higher","IoU":"higher","mIoU":null}},"counts":{"rows":23,"rows_with_code":21,"rows_with_paper_page":23,"rows_dated":23,"rows_using_additional_data":11},"rows":[{"rank_in_archive_order":1,"model":"MLCD-Seg-7B","metrics":{"Overall IoU":"79.9"},"uses_additional_data":true,"paper_date":"2024-07-24","paper":"/paper/multi-label-cluster-discrimination-for-visual","paper_url":"https://arxiv.org/abs/2407.17331v2","paper_title":"Multi-label Cluster Discrimination for Visual Representation Learning","code":"https://github.com/deepglint/unicom","n_code_links":1,"syntology":{"n_ran":7,"n_unverified":4,"n_samples":11,"n_pointer_only_licence":0}},{"rank_in_archive_order":2,"model":"HyperSeg","metrics":{"Overall IoU":"79.4"},"uses_additional_data":true,"paper_date":"2024-11-26","paper":"/paper/hyperseg-towards-universal-visual","paper_url":"https://arxiv.org/abs/2411.17606v2","paper_title":"HyperSeg: Towards Universal Visual Segmentation with Large Language Model","code":"https://github.com/congvvc/HyperSeg","n_code_links":1,"syntology":{"n_ran":7,"n_unverified":10,"n_samples":17,"n_pointer_only_licence":0}},{"rank_in_archive_order":3,"model":"UniLSeg-100","metrics":{"Overall IoU":"79.27"},"uses_additional_data":true,"paper_date":"2023-12-04","paper":"/paper/universal-segmentation-at-arbitrary","paper_url":"https://arxiv.org/abs/2312.01623v4","paper_title":"Universal Segmentation at Arbitrary Granularity with Language Instruction","code":"https://github.com/yongliu20/UniLSeg","n_code_links":2,"syntology":{"n_ran":13,"n_unverified":3,"n_samples":16,"n_pointer_only_licence":0}},{"rank_in_archive_order":4,"model":"UniLSeg-20","metrics":{"Overall IoU":"78.41"},"uses_additional_data":true,"paper_date":"2023-12-04","paper":"/paper/universal-segmentation-at-arbitrary","paper_url":"https://arxiv.org/abs/2312.01623v4","paper_title":"Universal Segmentation at Arbitrary Granularity with Language Instruction","code":"https://github.com/yongliu20/UniLSeg","n_code_links":2,"syntology":{"n_ran":13,"n_unverified":3,"n_samples":16,"n_pointer_only_licence":0}},{"rank_in_archive_order":5,"model":"EVF-SAM","metrics":{"Overall IoU":"78.2"},"uses_additional_data":true,"paper_date":"2024-06-28","paper":"/paper/evf-sam-early-vision-language-fusion-for-text","paper_url":"https://arxiv.org/abs/2406.20076v4","paper_title":"EVF-SAM: Early Vision-Language Fusion for Text-Prompted Segment Anything Model","code":"https://github.com/hustvl/evf-sam","n_code_links":1,"syntology":{"n_ran":3,"n_unverified":0,"n_samples":3,"n_pointer_only_licence":2}},{"rank_in_archive_order":6,"model":"SegAgent","metrics":{"Overall IoU":"75.11"},"uses_additional_data":false,"paper_date":"2025-03-11","paper":"/paper/segagent-exploring-pixel-understanding","paper_url":"https://arxiv.org/abs/2503.08625","paper_title":"SegAgent: Exploring Pixel Understanding Capabilities in MLLMs by Imitating Human Annotator Trajectories","code":"https://github.com/aim-uofa/SegAgent","n_code_links":1,"syntology":{"n_ran":1,"n_unverified":1,"n_samples":2,"n_pointer_only_licence":0}},{"rank_in_archive_order":7,"model":"DETRIS","metrics":{"Overall IoU":"74.6"},"uses_additional_data":false,"paper_date":"2025-01-15","paper":"/paper/densely-connected-parameter-efficient-tuning","paper_url":"https://arxiv.org/abs/2501.08580v1","paper_title":"Densely Connected Parameter-Efficient Tuning for Referring Image Segmentation","code":"https://github.com/jiaqihuang01/detris","n_code_links":1,"syntology":{"n_ran":7,"n_unverified":10,"n_samples":17,"n_pointer_only_licence":0}},{"rank_in_archive_order":8,"model":"C3VG","metrics":{"Overall IoU":"74.43"},"uses_additional_data":false,"paper_date":"2025-01-12","paper":"/paper/multi-task-visual-grounding-with-coarse-to","paper_url":"https://arxiv.org/abs/2501.06710v1","paper_title":"Multi-task Visual Grounding with Coarse-to-Fine Consistency Constraints","code":"https://github.com/dmmm1997/c3vg","n_code_links":1,"syntology":{"n_ran":5,"n_unverified":8,"n_samples":13,"n_pointer_only_licence":0}},{"rank_in_archive_order":9,"model":"GROUNDHOG","metrics":{"Overall IoU":"74.1"},"uses_additional_data":true,"paper_date":"2024-02-26","paper":"/paper/groundhog-grounding-large-language-models-to","paper_url":"https://arxiv.org/abs/2402.16846v2","paper_title":"GROUNDHOG: Grounding Large Language Models to Holistic Segmentation","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":10,"model":"GLEE-Pro","metrics":{"Overall IoU":"72.9"},"uses_additional_data":true,"paper_date":"2023-12-14","paper":"/paper/general-object-foundation-model-for-images","paper_url":"https://arxiv.org/abs/2312.09158v1","paper_title":"General Object Foundation Model for Images and Videos at Scale","code":"https://github.com/FoundationVision/GLEE","n_code_links":1,"syntology":{"n_ran":8,"n_unverified":5,"n_samples":13,"n_pointer_only_licence":0}},{"rank_in_archive_order":11,"model":"SafaRi-B","metrics":{"Overall IoU":"70.48"},"uses_additional_data":true,"paper_date":"2024-07-02","paper":"/paper/safari-adaptive-sequence-transformer-for","paper_url":"https://arxiv.org/abs/2407.02389v1","paper_title":"SafaRi:Adaptive Sequence Transformer for Weakly Supervised Referring Expression Segmentation","code":null,"n_code_links":0,"syntology":null},{"rank_in_archive_order":12,"model":"PolyFormer-L","metrics":{"Mean IoU":"71.15","Overall IoU":"69.2"},"uses_additional_data":true,"paper_date":"2023-02-14","paper":"/paper/polyformer-referring-image-segmentation-as","paper_url":"https://arxiv.org/abs/2302.07387v2","paper_title":"PolyFormer: Referring Image Segmentation as Sequential Polygon Generation","code":"https://github.com/amazon-science/polygon-transformer","n_code_links":1,"syntology":{"n_ran":5,"n_unverified":1,"n_samples":6,"n_pointer_only_licence":6}},{"rank_in_archive_order":13,"model":"MaskRIS (Swin-B, combined DB)","metrics":{"Overall IoU":"69.12"},"uses_additional_data":false,"paper_date":"2024-11-28","paper":"/paper/maskris-semantic-distortion-aware-data","paper_url":"https://arxiv.org/abs/2411.19067v1","paper_title":"MaskRIS: Semantic Distortion-aware Data Augmentation for Referring Image Segmentation","code":"https://github.com/naver-ai/maskris","n_code_links":1,"syntology":null},{"rank_in_archive_order":14,"model":"PolyFormer-B","metrics":{"Mean IoU":"69.36","Overall IoU":"67.76"},"uses_additional_data":true,"paper_date":"2023-02-14","paper":"/paper/polyformer-referring-image-segmentation-as","paper_url":"https://arxiv.org/abs/2302.07387v2","paper_title":"PolyFormer: Referring Image Segmentation as Sequential Polygon Generation","code":"https://github.com/amazon-science/polygon-transformer","n_code_links":1,"syntology":{"n_ran":5,"n_unverified":1,"n_samples":6,"n_pointer_only_licence":6}},{"rank_in_archive_order":15,"model":"MaskRIS (Swin-B)","metrics":{"Mean IoU":"69.31","Overall IoU":"65.55"},"uses_additional_data":false,"paper_date":"2024-11-28","paper":"/paper/maskris-semantic-distortion-aware-data","paper_url":"https://arxiv.org/abs/2411.19067v1","paper_title":"MaskRIS: Semantic Distortion-aware Data Augmentation for Referring Image Segmentation","code":"https://github.com/naver-ai/maskris","n_code_links":1,"syntology":null},{"rank_in_archive_order":16,"model":"MagNet","metrics":{"Overall IoU":"65.36"},"uses_additional_data":false,"paper_date":"2023-12-19","paper":"/paper/mask-grounding-for-referring-image","paper_url":"https://arxiv.org/abs/2312.12198v2","paper_title":"Mask Grounding for Referring Image Segmentation","code":"https://github.com/yxchng/mask-grounding","n_code_links":1,"syntology":{"n_ran":10,"n_unverified":3,"n_samples":13,"n_pointer_only_licence":13}},{"rank_in_archive_order":17,"model":"X-Decoder (Davit-d5)","metrics":{"Overall IoU":"64.6"},"uses_additional_data":true,"paper_date":"2022-12-21","paper":"/paper/generalized-decoding-for-pixel-image-and","paper_url":"https://arxiv.org/abs/2212.11270v1","paper_title":"Generalized Decoding for Pixel, Image, and Language","code":"https://github.com/microsoft/X-Decoder","n_code_links":1,"syntology":null},{"rank_in_archive_order":18,"model":"VLT (Swin-B)","metrics":{"Overall IoU":"63.49"},"uses_additional_data":false,"paper_date":"2022-10-28","paper":"/paper/vlt-vision-language-transformer-and-query","paper_url":"https://arxiv.org/abs/2210.15871v1","paper_title":"VLT: Vision-Language Transformer and Query Generation for Referring Segmentation","code":"https://github.com/henghuiding/Vision-Language-Transformer","n_code_links":1,"syntology":{"n_ran":3,"n_unverified":3,"n_samples":6,"n_pointer_only_licence":0}},{"rank_in_archive_order":19,"model":"LAVT","metrics":{"Overall IoU":"61.24"},"uses_additional_data":false,"paper_date":"2021-12-04","paper":"/paper/lavt-language-aware-vision-transformer-for","paper_url":"https://arxiv.org/abs/2112.02244v2","paper_title":"LAVT: Language-Aware Vision Transformer for Referring Image Segmentation","code":"https://github.com/yz93/lavt-ris","n_code_links":1,"syntology":null},{"rank_in_archive_order":20,"model":"VLT (Darknet53)","metrics":{"Overall IoU":"52.99"},"uses_additional_data":false,"paper_date":"2021-08-12","paper":"/paper/vision-language-transformer-and-query","paper_url":"https://arxiv.org/abs/2108.05565v1","paper_title":"Vision-Language Transformer and Query Generation for Referring Segmentation","code":"https://github.com/henghuiding/Vision-Language-Transformer","n_code_links":1,"syntology":null},{"rank_in_archive_order":21,"model":"SHNet","metrics":{"Overall IoU":"49.90"},"uses_additional_data":false,"paper_date":"2021-04-21","paper":"/paper/comprehensive-multi-modal-interactions-for","paper_url":"https://arxiv.org/abs/2104.10412v4","paper_title":"Comprehensive Multi-Modal Interactions for Referring Image Segmentation","code":"https://github.com/kanji95/SHNET","n_code_links":1,"syntology":null},{"rank_in_archive_order":22,"model":"DeRIS-L","metrics":{"Mean IoU":"80.01"},"uses_additional_data":false,"paper_date":"2025-07-02","paper":"/paper/deris-decoupling-perception-and-cognition-for","paper_url":"https://arxiv.org/abs/2507.01738v1","paper_title":"DeRIS: Decoupling Perception and Cognition for Enhanced Referring Image Segmentation through Loopback Synergy","code":"https://github.com/Dmmm1997/DeRIS","n_code_links":1,"syntology":null},{"rank_in_archive_order":23,"model":"VATEX","metrics":{"IoU":"0.7554","mIoU":"69.73"},"uses_additional_data":false,"paper_date":"2024-04-12","paper":"/paper/improving-referring-image-segmentation-using","paper_url":"https://arxiv.org/abs/2404.08590v2","paper_title":"Vision-Aware Text Features in Referring Image Segmentation: From Object Understanding to Context Understanding","code":"https://github.com/nero1342/VATEX","n_code_links":1,"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 6,885 of the 9,623 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":9623,"papers_checked":6885,"papers_extracted_not_yet_verified":0,"boards_without_verdict":2,"papers_not_yet_extracted":2737},"order":"newest first by month (arXiv date, else the arXiv-id month), then arXiv id descending","columns":[],"entries":[]},"syntology":{"read_at":"2026-09-25T09:33:49+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":13,"rows_with_any_sample_ran":13,"distinct_papers_with_graph_line":11,"distinct_papers_with_any_sample_ran":11,"samples_over_distinct_papers":{"n_ran":69,"n_unverified":48,"n_samples":117,"n_pointer_only_licence":21,"note":"each paper (arXiv id) counted once, however many rows it is behind; this is the page-level figure"},"samples_row_weighted":{"n_ran":87,"n_unverified":52,"n_samples":139,"n_pointer_only_licence":27,"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"}}}