{"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/fully-convolutional-networks-for-panoptic","title":"Fully Convolutional Networks for Panoptic Segmentation","arxiv_id":"2012.00720","date":"2020-12-01","proceeding":"CVPR 2021 1","authors":["Yanwei Li","Hengshuang Zhao","Xiaojuan Qi","LiWei Wang","Zeming Li","Jian Sun","Jiaya Jia"],"abstract":"In this paper, we present a conceptually simple, strong, and efficient framework for panoptic segmentation, called Panoptic FCN. Our approach aims to represent and predict foreground things and background stuff in a unified fully convolutional pipeline. In particular, Panoptic FCN encodes each object instance or stuff category into a specific kernel weight with the proposed kernel generator and produces the prediction by convolving the high-resolution feature directly. With this approach, instance-aware and semantically consistent properties for things and stuff can be respectively satisfied in a simple generate-kernel-then-segment workflow. Without extra boxes for localization or instance separation, the proposed approach outperforms previous box-based and -free models with high efficiency on COCO, Cityscapes, and Mapillary Vistas datasets with single scale input. Our code is made publicly available at https://github.com/Jia-Research-Lab/PanopticFCN.","url_abs":"https://arxiv.org/abs/2012.00720v2","url_pdf":"https://arxiv.org/pdf/2012.00720v2.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":"fully-convolutional-networks-for-panoptic","repo_url":"https://github.com/Jia-Research-Lab/PanopticFCN","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"unanswered"}},{"paper_slug":"fully-convolutional-networks-for-panoptic","repo_url":"https://github.com/yanwei-li/PanopticFCN","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"unanswered"}},{"paper_slug":"fully-convolutional-networks-for-panoptic","repo_url":"https://github.com/DdeGeus/PanopticFCN-IBS","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"unanswered"}},{"paper_slug":"fully-convolutional-networks-for-panoptic","repo_url":"https://github.com/Jia-Research-Lab/MSAD","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"fully-convolutional-networks-for-panoptic","repo_url":"https://github.com/dvlab-research/msad","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}},{"paper_slug":"fully-convolutional-networks-for-panoptic","repo_url":"https://github.com/dvlab-research/panopticfcn","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"Apache-2.0"}}],"tasks":[{"task_slug":"panoptic-segmentation","task_name":"Panoptic Segmentation"},{"task_slug":"segmentation","task_name":"Segmentation"}],"methods":[{"method_slug":"convolution","method_name":"Convolution"},{"method_slug":"fcn","method_name":"FCN"},{"method_slug":"max-pooling","method_name":"Max Pooling"}],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/panoptic-segmentation-on-coco-minival","task":"Panoptic Segmentation","dataset":"COCO minival","model":"Panoptic FCN* (ResNet-50-FPN)","rank_in_archive_order":26,"of":31,"metrics":{"PQ":"44.3","PQst":"35.6","PQth":"50","RQ":"53","RQst":"43.5","RQth":"59.3","SQ":"80.7","SQst":"76.7","SQth":"83.4"},"uses_additional_data":false},{"leaderboard":"/sota/panoptic-segmentation-on-coco-minival","task":"Panoptic Segmentation","dataset":"COCO minival","model":"Panoptic FCN* (Swin-L, single-scale)","rank_in_archive_order":31,"of":31,"metrics":{"PQth":"58.5 ","RQ":" 61.6","RQst":"51.1","RQth":"68.6 ","SQ":"83.2","SQst":"81.1","SQth":"84.6"},"uses_additional_data":false},{"leaderboard":"/sota/panoptic-segmentation-on-coco-test-dev","task":"Panoptic Segmentation","dataset":"COCO test-dev","model":"Panoptic FCN* (Swin-L)","rank_in_archive_order":10,"of":38,"metrics":{"PQ":"52.7","PQth":" 59.4"},"uses_additional_data":false},{"leaderboard":"/sota/panoptic-segmentation-on-coco-test-dev","task":"Panoptic Segmentation","dataset":"COCO test-dev","model":"Panoptic FCN*++ (DCN-101-FPN)","rank_in_archive_order":21,"of":38,"metrics":{"PQ":"47.5","PQst":"38.2","PQth":"53.7"},"uses_additional_data":false},{"leaderboard":"/sota/panoptic-segmentation-on-cityscapes-val","task":"Panoptic Segmentation","dataset":"Cityscapes val","model":"Panoptic FCN* (ResNet-FPN)","rank_in_archive_order":23,"of":37,"metrics":{"PQ":"61.4","PQth":"54.8"},"uses_additional_data":false},{"leaderboard":"/sota/panoptic-segmentation-on-cityscapes-val","task":"Panoptic Segmentation","dataset":"Cityscapes val","model":"Panoptic FCN* (Swin-L, Cityscapes-fine)","rank_in_archive_order":35,"of":37,"metrics":{"PQst":" 70.6","PQth":"59.5"},"uses_additional_data":false},{"leaderboard":"/sota/panoptic-segmentation-on-cityscapes-val","task":"Panoptic Segmentation","dataset":"Cityscapes val","model":"Panoptic FCN* (ResNet-50-FPN)","rank_in_archive_order":36,"of":37,"metrics":{"PQst":"66.6"},"uses_additional_data":false},{"leaderboard":"/sota/panoptic-segmentation-on-mapillary-val","task":"Panoptic Segmentation","dataset":"Mapillary val","model":"Panoptic FCN* (Swin-L, single-scale)","rank_in_archive_order":3,"of":13,"metrics":{"PQ":"45.7","PQst":"52.1","PQth":"40.8"},"uses_additional_data":false},{"leaderboard":"/sota/panoptic-segmentation-on-mapillary-val","task":"Panoptic Segmentation","dataset":"Mapillary val","model":"Panoptic FCN* (ResNet-FPN)","rank_in_archive_order":10,"of":13,"metrics":{"PQ":"36.9","PQth":"32.9"},"uses_additional_data":false},{"leaderboard":"/sota/panoptic-segmentation-on-mapillary-val","task":"Panoptic Segmentation","dataset":"Mapillary val","model":"Panoptic FCN* (ResNet-50-FPN)","rank_in_archive_order":13,"of":13,"metrics":{"PQst":"42.3"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=2012.00720","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}