{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/masked-attention-mask-transformer-for","title":"Masked-attention Mask Transformer for Universal Image Segmentation","arxiv_id":"2112.01527","date":"2021-12-02","proceeding":"CVPR 2022 1","authors":["Bowen Cheng","Ishan Misra","Alexander G. Schwing","Alexander Kirillov","Rohit Girdhar"],"abstract":"Image segmentation is about grouping pixels with different semantics, e.g., category or instance membership, where each choice of semantics defines a task. While only the semantics of each task differ, current research focuses on designing specialized architectures for each task. We present Masked-attention Mask Transformer (Mask2Former), a new architecture capable of addressing any image segmentation task (panoptic, instance or semantic). Its key components include masked attention, which extracts localized features by constraining cross-attention within predicted mask regions. In addition to reducing the research effort by at least three times, it outperforms the best specialized architectures by a significant margin on four popular datasets. Most notably, Mask2Former sets a new state-of-the-art for panoptic segmentation (57.8 PQ on COCO), instance segmentation (50.1 AP on COCO) and semantic segmentation (57.7 mIoU on ADE20K).","url_abs":"https://arxiv.org/abs/2112.01527v3","url_pdf":"https://arxiv.org/pdf/2112.01527v3.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"masked-attention-mask-transformer-for","repo_url":"https://github.com/facebookresearch/Mask2Former","is_official":1,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"MIT"}},{"paper_slug":"masked-attention-mask-transformer-for","repo_url":"https://github.com/DdeGeus/Mask2Former-IBS","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"MIT"}},{"paper_slug":"masked-attention-mask-transformer-for","repo_url":"https://github.com/huggingface/transformers","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"masked-attention-mask-transformer-for","repo_url":"https://github.com/nihalsid/mask2former","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"MIT"}},{"paper_slug":"masked-attention-mask-transformer-for","repo_url":"https://github.com/MindSpore-scientific/code-7/tree/main/Mask2Former","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"none","reach":null},{"paper_slug":"masked-attention-mask-transformer-for","repo_url":"https://github.com/alibaba/EasyCV","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"pytorch","reach":null},{"paper_slug":"masked-attention-mask-transformer-for","repo_url":"https://github.com/open-mmlab/mmdetection","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"pytorch","reach":{"status":"ok","spdx":"Apache-2.0"}}],"tasks":[{"task_slug":"2d-semantic-segmentation","task_name":"2D Semantic Segmentation"},{"task_slug":"image-segmentation","task_name":"Image Segmentation"},{"task_slug":"instance-segmentation","task_name":"Instance Segmentation"},{"task_slug":"panoptic-segmentation","task_name":"Panoptic Segmentation"},{"task_slug":"segmentation","task_name":"Segmentation"},{"task_slug":"semantic-segmentation","task_name":"Semantic Segmentation"},{"task_slug":"universal-segmentation","task_name":"Universal Segmentation"}],"methods":[{"method_slug":"absolute-position-encodings","method_name":"Absolute Position Encodings"},{"method_slug":"adam","method_name":"Adam"},{"method_slug":"attention","method_name":"Attention"},{"method_slug":"bpe","method_name":"BPE"},{"method_slug":"dense-connections","method_name":"Dense Connections"},{"method_slug":"dropout","method_name":"Dropout"},{"method_slug":"label-smoothing","method_name":"Label Smoothing"},{"method_slug":"layer-normalization","method_name":"Layer Normalization"},{"method_slug":"linear-layer","method_name":"Linear Layer"},{"method_slug":"multi-head-attention","method_name":"Multi-Head Attention"},{"method_slug":"position-wise-feed-forward-layer","method_name":"Position-Wise Feed-Forward Layer"},{"method_slug":"residual-connection","method_name":"Residual Connection"},{"method_slug":"softmax","method_name":"Softmax"},{"method_slug":"transformer","method_name":"Transformer"}],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/2d-semantic-segmentation-on-wildscenes","task":"2D Semantic Segmentation","dataset":"WildScenes","model":"Mask2Former (Swin-L)","rank_in_archive_order":1,"of":5,"metrics":{"mIoU":"47.85"},"uses_additional_data":true},{"leaderboard":"/sota/2d-semantic-segmentation-on-wildscenes","task":"2D Semantic Segmentation","dataset":"WildScenes","model":"Mask2Former (ResNet-50)","rank_in_archive_order":3,"of":5,"metrics":{"mIoU":"43.71"},"uses_additional_data":true},{"leaderboard":"/sota/instance-segmentation-on-ade20k-val","task":"Instance Segmentation","dataset":"ADE20K val","model":"Mask2Former (Swin-L, single-scale)","rank_in_archive_order":11,"of":14,"metrics":{"AP":"34.9","APL":"54.7","APM":"40","APS":"16.3"},"uses_additional_data":false},{"leaderboard":"/sota/instance-segmentation-on-ade20k-val","task":"Instance Segmentation","dataset":"ADE20K val","model":"Mask2Former (Swin-L + FAPN)","rank_in_archive_order":12,"of":14,"metrics":{"AP":"33.4","APL":"54.6","APM":"37.6","APS":"14.6"},"uses_additional_data":false},{"leaderboard":"/sota/instance-segmentation-on-ade20k-val","task":"Instance Segmentation","dataset":"ADE20K val","model":"Mask2Former (ResNet50)","rank_in_archive_order":13,"of":14,"metrics":{"AP":"26.4","APS":"10.4"},"uses_additional_data":false},{"leaderboard":"/sota/instance-segmentation-on-ade20k-val","task":"Instance Segmentation","dataset":"ADE20K val","model":"Mask2Former (ResNet-50)","rank_in_archive_order":14,"of":14,"metrics":{"APL":"43.1","APM":"28.9"},"uses_additional_data":false},{"leaderboard":"/sota/instance-segmentation-on-coco-minival","task":"Instance Segmentation","dataset":"COCO minival","model":"Mask2Former (Swin-L)","rank_in_archive_order":26,"of":93,"metrics":{"mask AP":"50.1"},"uses_additional_data":false},{"leaderboard":"/sota/instance-segmentation-on-coco","task":"Instance Segmentation","dataset":"COCO test-dev","model":"Mask2Former (Swin-L, single scale)","rank_in_archive_order":21,"of":112,"metrics":{"AP50":"74.9","AP75":"54.9","APL":"71.2","APM":"53.8","APS":"29.1","mask AP":"50.5"},"uses_additional_data":false},{"leaderboard":"/sota/instance-segmentation-on-coco-val-panoptic","task":"Instance Segmentation","dataset":"COCO val (panoptic labels)","model":"Mask2Former (Swin-L, single-scale)","rank_in_archive_order":3,"of":4,"metrics":{"AP":"49.1"},"uses_additional_data":false},{"leaderboard":"/sota/instance-segmentation-on-cityscapes-val","task":"Instance Segmentation","dataset":"Cityscapes val","model":"Mask2Former (Swin-L, single-scale)","rank_in_archive_order":9,"of":17,"metrics":{"mask AP":"43.7"},"uses_additional_data":false},{"leaderboard":"/sota/instance-segmentation-on-cityscapes-val","task":"Instance Segmentation","dataset":"Cityscapes val","model":"Mask2Former (Swin-B)","rank_in_archive_order":10,"of":17,"metrics":{"mask AP":"42"},"uses_additional_data":false},{"leaderboard":"/sota/instance-segmentation-on-cityscapes-val","task":"Instance Segmentation","dataset":"Cityscapes val","model":"Mask2Former (Swin-S)","rank_in_archive_order":11,"of":17,"metrics":{"mask AP":"41.8"},"uses_additional_data":false},{"leaderboard":"/sota/instance-segmentation-on-cityscapes-val","task":"Instance Segmentation","dataset":"Cityscapes val","model":"Mask2Former (Swin-T)","rank_in_archive_order":13,"of":17,"metrics":{"mask AP":"39.7"},"uses_additional_data":false},{"leaderboard":"/sota/instance-segmentation-on-cityscapes-val","task":"Instance Segmentation","dataset":"Cityscapes val","model":"Mask2Former (ResNet-101)","rank_in_archive_order":14,"of":17,"metrics":{"mask AP":"38.5"},"uses_additional_data":false},{"leaderboard":"/sota/instance-segmentation-on-cityscapes-val","task":"Instance Segmentation","dataset":"Cityscapes val","model":"Mask2Former (ResNet-50)","rank_in_archive_order":15,"of":17,"metrics":{"mask AP":"37.4"},"uses_additional_data":false},{"leaderboard":"/sota/panoptic-segmentation-on-ade20k-val","task":"Panoptic Segmentation","dataset":"ADE20K val","model":"Mask2Former (Swin-L)","rank_in_archive_order":18,"of":25,"metrics":{"AP":"34.2","PQ":"48.1","mIoU":"54.5"},"uses_additional_data":false},{"leaderboard":"/sota/panoptic-segmentation-on-ade20k-val","task":"Panoptic Segmentation","dataset":"ADE20K val","model":"Mask2Former (Swin-L + FAPN, 640x640)","rank_in_archive_order":19,"of":25,"metrics":{"AP":"33.2","PQ":"46.2","mIoU":"55.4"},"uses_additional_data":false},{"leaderboard":"/sota/panoptic-segmentation-on-ade20k-val","task":"Panoptic Segmentation","dataset":"ADE20K val","model":"Mask2Former (ResNet-50, 640x640)","rank_in_archive_order":22,"of":25,"metrics":{"PQ":"39.7"},"uses_additional_data":false},{"leaderboard":"/sota/panoptic-segmentation-on-ade20k-val","task":"Panoptic Segmentation","dataset":"ADE20K val","model":"Panoptic-DeepLab (SwideRNet)","rank_in_archive_order":23,"of":25,"metrics":{"PQ":"37.9","mIoU":"50"},"uses_additional_data":false},{"leaderboard":"/sota/panoptic-segmentation-on-ade20k-val","task":"Panoptic Segmentation","dataset":"ADE20K val","model":"Mask2Former (ResNet-50, 640x640)","rank_in_archive_order":25,"of":25,"metrics":{"AP":"26.5","mIoU":"46.1"},"uses_additional_data":false},{"leaderboard":"/sota/panoptic-segmentation-on-coco-minival","task":"Panoptic Segmentation","dataset":"COCO minival","model":"Mask2Former (single-scale)","rank_in_archive_order":18,"of":31,"metrics":{"AP":"48.6","PQ":"57.8","PQst":"48.1","PQth":"64.2"},"uses_additional_data":false},{"leaderboard":"/sota/panoptic-segmentation-on-coco-test-dev","task":"Panoptic Segmentation","dataset":"COCO test-dev","model":"Mask2Former (Swin-L)","rank_in_archive_order":3,"of":38,"metrics":{"PQ":"58.3","PQst":"48.1","PQth":"65.1"},"uses_additional_data":false},{"leaderboard":"/sota/panoptic-segmentation-on-cityscapes-val","task":"Panoptic Segmentation","dataset":"Cityscapes val","model":"Mask2Former (Swin-L)","rank_in_archive_order":15,"of":37,"metrics":{"AP":"43.6","PQ":"66.6","mIoU":"82.9"},"uses_additional_data":false},{"leaderboard":"/sota/semantic-segmentation-on-ade20k","task":"Semantic Segmentation","dataset":"ADE20K","model":"Mask2Former (SwinL-FaPN)","rank_in_archive_order":27,"of":235,"metrics":{"Validation mIoU":"57.7"},"uses_additional_data":false},{"leaderboard":"/sota/semantic-segmentation-on-ade20k","task":"Semantic Segmentation","dataset":"ADE20K","model":"Mask2Former (SwinL)","rank_in_archive_order":33,"of":235,"metrics":{"Validation mIoU":"57.3"},"uses_additional_data":false},{"leaderboard":"/sota/semantic-segmentation-on-ade20k","task":"Semantic Segmentation","dataset":"ADE20K","model":"Mask2Former (Swin-L-FaPN)","rank_in_archive_order":40,"of":235,"metrics":{"Validation mIoU":"56.4"},"uses_additional_data":false},{"leaderboard":"/sota/semantic-segmentation-on-ade20k","task":"Semantic Segmentation","dataset":"ADE20K","model":"Mask2Former(Swin-B)","rank_in_archive_order":50,"of":235,"metrics":{"Validation mIoU":"55.1"},"uses_additional_data":false},{"leaderboard":"/sota/semantic-segmentation-on-ade20k-val","task":"Semantic Segmentation","dataset":"ADE20K val","model":"Mask2Former (Swin-L-FaPN, multiscale)","rank_in_archive_order":18,"of":95,"metrics":{"mIoU":"57.7"},"uses_additional_data":false},{"leaderboard":"/sota/semantic-segmentation-on-ade20k-val","task":"Semantic Segmentation","dataset":"ADE20K val","model":"Mask2Former (Swin-L-FaPN)","rank_in_archive_order":25,"of":95,"metrics":{"mIoU":"56.4"},"uses_additional_data":false},{"leaderboard":"/sota/semantic-segmentation-on-coco-1","task":"Semantic Segmentation","dataset":"COCO (Common Objects in Context)","model":"Mask2Former (Swin-L, single-scale)","rank_in_archive_order":7,"of":9,"metrics":{"mIoU":"67.4"},"uses_additional_data":false},{"leaderboard":"/sota/semantic-segmentation-on-coco-1","task":"Semantic Segmentation","dataset":"COCO (Common Objects in Context)","model":"MaskFormer (Swin-L, single-scale)","rank_in_archive_order":8,"of":9,"metrics":{"mIoU":"64.8"},"uses_additional_data":false},{"leaderboard":"/sota/semantic-segmentation-on-cityscapes-val","task":"Semantic Segmentation","dataset":"Cityscapes val","model":"Mask2Former (Swin-L)","rank_in_archive_order":19,"of":99,"metrics":{"mIoU":"84.3"},"uses_additional_data":false},{"leaderboard":"/sota/semantic-segmentation-on-fine-grained-grass","task":"Semantic Segmentation","dataset":"Fine-Grained Grass Segmentation Dataset","model":"Mask2Former","rank_in_archive_order":10,"of":10,"metrics":{"mIoU":"44.93"},"uses_additional_data":false},{"leaderboard":"/sota/semantic-segmentation-on-mapillary-val","task":"Semantic Segmentation","dataset":"Mapillary val","model":"Mask2Former (Swin-L, multiscale)","rank_in_archive_order":3,"of":8,"metrics":{"mIoU":"64.7"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=2112.01527","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2112.01527"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-24T18:15:14+00:00","read_at_is":"when the build read Syntology's graph, not when any sample ran","claim":"Per-sample execution status on synthesized fixtures; not a correctness claim about the paper. Samples come from repositories linked to the paper, official or community; repo_kind says which.","repos":[{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/huggingface/transformers","reach":null},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/MindSpore-scientific/code-7/tree/main/Mask2Former","reach":null},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/open-mmlab/mmdetection","reach":{"status":"ok","spdx":"Apache-2.0"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/alibaba/EasyCV","reach":null},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/DdeGeus/Mask2Former-IBS","reach":{"status":"ok","spdx":"MIT"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/facebookresearch/Mask2Former","reach":{"status":"ok","spdx":"MIT"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/nihalsid/mask2former","reach":{"status":"ok","spdx":"MIT"}}],"summary":{"ran_fixture":1,"ran_draft_wrong":1,"unverified":6},"by_repo_kind":{"listed":{"samples":8,"ran":2,"repositories":1}},"repo_kind_vocabulary":{"official":"The archive marks this repository official for the paper","named_in_paper":"The archive records that the paper mentions this repository; it is not marked official","listed":"In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper","found_in_text":"Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted","community":"Not in the archive's code links for this paper; a community repository Syntology harvested"},"n_pointer_only_for_licence":0,"samples":[{"code_sha256_prefix":"f9fd6241d935f07b","entry":"window_partition","repo":"DdeGeus/Mask2Former-IBS","repo_kind":"listed","path":"mask2former/modeling/backbone/swin.py","file_url":"https://github.com/DdeGeus/Mask2Former-IBS/blob/HEAD/mask2former/modeling/backbone/swin.py","link_basis":"harvester_set","language":"python","status":"ran_fixture","verification_level":1,"contract_check":"RAISES","metamorphic_tier":"invariant","behaviour_fingerprint":true,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"f9fd6241d935f07b"}},{"code_sha256_prefix":"fb32094c6dbece71","entry":"window_reverse","repo":"DdeGeus/Mask2Former-IBS","repo_kind":"listed","path":"mask2former/modeling/backbone/swin.py","file_url":"https://github.com/DdeGeus/Mask2Former-IBS/blob/HEAD/mask2former/modeling/backbone/swin.py","link_basis":"harvester_set","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"MISDECLARED","metamorphic_tier":"invariant","behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"fb32094c6dbece71"}},{"code_sha256_prefix":"bc2cb481a75c370d","entry":"batch_dice_loss","repo":"DdeGeus/Mask2Former-IBS","repo_kind":"listed","path":"mask2former/modeling/matcher.py","file_url":"https://github.com/DdeGeus/Mask2Former-IBS/blob/HEAD/mask2former/modeling/matcher.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"bc2cb481a75c370d"}},{"code_sha256_prefix":"1edd24985036b0bf","entry":"batch_sigmoid_ce_loss","repo":"DdeGeus/Mask2Former-IBS","repo_kind":"listed","path":"mask2former/modeling/matcher.py","file_url":"https://github.com/DdeGeus/Mask2Former-IBS/blob/HEAD/mask2former/modeling/matcher.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"1edd24985036b0bf"}},{"code_sha256_prefix":"2dcb8123d89bb1ff","entry":"calculate_uncertainty","repo":"DdeGeus/Mask2Former-IBS","repo_kind":"listed","path":"mask2former_video/modeling/criterion.py","file_url":"https://github.com/DdeGeus/Mask2Former-IBS/blob/HEAD/mask2former_video/modeling/criterion.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"2dcb8123d89bb1ff"}},{"code_sha256_prefix":"89f75e54ff128be0","entry":"dice_loss","repo":"DdeGeus/Mask2Former-IBS","repo_kind":"listed","path":"mask2former/modeling/criterion.py","file_url":"https://github.com/DdeGeus/Mask2Former-IBS/blob/HEAD/mask2former/modeling/criterion.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"89f75e54ff128be0"}},{"code_sha256_prefix":"2bb4c1887f16e1ba","entry":"focal_loss","repo":"DdeGeus/Mask2Former-IBS","repo_kind":"listed","path":"mask2former/modeling/criterion.py","file_url":"https://github.com/DdeGeus/Mask2Former-IBS/blob/HEAD/mask2former/modeling/criterion.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"2bb4c1887f16e1ba"}},{"code_sha256_prefix":"d0c61e8dba511aa3","entry":"sigmoid_ce_loss","repo":"DdeGeus/Mask2Former-IBS","repo_kind":"listed","path":"mask2former/modeling/criterion.py","file_url":"https://github.com/DdeGeus/Mask2Former-IBS/blob/HEAD/mask2former/modeling/criterion.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"d0c61e8dba511aa3"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}