Hybrid Task Cascade for Instance Segmentation
Cascade is a classic yet powerful architecture that has boosted performance on various tasks. However, how to introduce cascade to instance segmentation remains an open question. A simple combination of Cascade R-CNN and Mask R-CNN only brings limited gain. In exploring a more effective approach, we...
Saved in:
| Published in: | Proceedings (IEEE Computer Society Conference on Computer Vision and Pattern Recognition. Online) pp. 4969 - 4978 |
|---|---|
| Main Authors: | , , , , , , , , , , , |
| Format: | Conference Proceeding |
| Language: | English |
| Published: |
IEEE
01.06.2019
|
| Subjects: | |
| ISSN: | 1063-6919 |
| Online Access: | Get full text |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Abstract | Cascade is a classic yet powerful architecture that has boosted performance on various tasks. However, how to introduce cascade to instance segmentation remains an open question. A simple combination of Cascade R-CNN and Mask R-CNN only brings limited gain. In exploring a more effective approach, we find that the key to a successful instance segmentation cascade is to fully leverage the reciprocal relationship between detection and segmentation. In this work, we propose a new framework, Hybrid Task Cascade (HTC), which differs in two important aspects: (1) instead of performing cascaded refinement on these two tasks separately, it interweaves them for a joint multi-stage processing; (2) it adopts a fully convolutional branch to provide spatial context, which can help distinguishing hard foreground from cluttered background. Overall, this framework can learn more discriminative features progressively while integrating complementary features together in each stage. Without bells and whistles, a single HTC obtains 38.4% and 1.5% improvement over a strong Cascade Mask R-CNN baseline on MSCOCO dataset. Moreover, our overall system achieves 48.6 mask AP on the test-challenge split, ranking 1st in the COCO 2018 Challenge Object Detection Task. Code is available at https://github.com/open-mmlab/mmdetection. |
|---|---|
| AbstractList | Cascade is a classic yet powerful architecture that has boosted performance on various tasks. However, how to introduce cascade to instance segmentation remains an open question. A simple combination of Cascade R-CNN and Mask R-CNN only brings limited gain. In exploring a more effective approach, we find that the key to a successful instance segmentation cascade is to fully leverage the reciprocal relationship between detection and segmentation. In this work, we propose a new framework, Hybrid Task Cascade (HTC), which differs in two important aspects: (1) instead of performing cascaded refinement on these two tasks separately, it interweaves them for a joint multi-stage processing; (2) it adopts a fully convolutional branch to provide spatial context, which can help distinguishing hard foreground from cluttered background. Overall, this framework can learn more discriminative features progressively while integrating complementary features together in each stage. Without bells and whistles, a single HTC obtains 38.4% and 1.5% improvement over a strong Cascade Mask R-CNN baseline on MSCOCO dataset. Moreover, our overall system achieves 48.6 mask AP on the test-challenge split, ranking 1st in the COCO 2018 Challenge Object Detection Task. Code is available at https://github.com/open-mmlab/mmdetection. |
| Author | Feng, Wansen Pang, Jiangmiao Ouyang, Wanli Chen, Kai Li, Xiaoxiao Shi, Jianping Liu, Ziwei Lin, Dahua Loy, Chen Change Wang, Jiaqi Xiong, Yu Sun, Shuyang |
| Author_xml | – sequence: 1 givenname: Kai surname: Chen fullname: Chen, Kai organization: The Chinese Univ. of Hong Kong – sequence: 2 givenname: Jiangmiao surname: Pang fullname: Pang, Jiangmiao organization: Zhejiang Univ – sequence: 3 givenname: Jiaqi surname: Wang fullname: Wang, Jiaqi organization: The Chinese Univ. of Hong Kong – sequence: 4 givenname: Yu surname: Xiong fullname: Xiong, Yu organization: The Chinese Univ. of HK – sequence: 5 givenname: Xiaoxiao surname: Li fullname: Li, Xiaoxiao organization: The Chinese Univ. of Hong Kong – sequence: 6 givenname: Shuyang surname: Sun fullname: Sun, Shuyang organization: The Univ. of Sydney – sequence: 7 givenname: Wansen surname: Feng fullname: Feng, Wansen organization: Lille Univ – sequence: 8 givenname: Ziwei surname: Liu fullname: Liu, Ziwei organization: The Chinese Univ. of Hong Kong – sequence: 9 givenname: Jianping surname: Shi fullname: Shi, Jianping organization: Sensetime Group Limited – sequence: 10 givenname: Wanli surname: Ouyang fullname: Ouyang, Wanli organization: The Univ. of Sydney – sequence: 11 givenname: Chen Change surname: Loy fullname: Loy, Chen Change organization: Nanyang Technological Univ – sequence: 12 givenname: Dahua surname: Lin fullname: Lin, Dahua organization: The Chinese Univ. of Hong Kong |
| BookMark | eNotzLFOwzAQgGGDQKKUzAwseYGUOzuxfSOKoK1UCQSFtbraFxSgDoqz9O1BgukfPum_VGdpSKLUNcICEei2fXt6XmhAWgA0iCeqIOfRaY9Gk_GnaoZgTWUJ6UIVOX8AgNGIlvxMVavjfuxjueX8WbacA0cpu2Es1ylPnIKUL_J-kDTx1A_pSp13_JWl-O9cvT7cb9tVtXlcrtu7TdVrMFOFXeRAXaAIQSJrjMGI81B3IUb20Tj0iA4liG6MDd4xN3b_a2Qdx9rM1c3ftxeR3ffYH3g87jw1NVprfgDF70V- |
| ContentType | Conference Proceeding |
| DBID | 6IE 6IH CBEJK RIE RIO |
| DOI | 10.1109/CVPR.2019.00511 |
| DatabaseName | IEEE Electronic Library (IEL) Conference Proceedings IEEE Proceedings Order Plan (POP) 1998-present by volume IEEE Xplore All Conference Proceedings IEEE Electronic Library (IEL) IEEE Proceedings Order Plans (POP) 1998-present |
| DatabaseTitleList | |
| Database_xml | – sequence: 1 dbid: RIE name: IEEE Electronic Library (IEL) url: https://ieeexplore.ieee.org/ sourceTypes: Publisher |
| DeliveryMethod | fulltext_linktorsrc |
| Discipline | Applied Sciences |
| EISBN | 9781728132938 1728132932 |
| EISSN | 1063-6919 |
| EndPage | 4978 |
| ExternalDocumentID | 8954166 |
| Genre | orig-research |
| GroupedDBID | 6IE 6IH 6IL 6IN AAWTH ABLEC ADZIZ ALMA_UNASSIGNED_HOLDINGS BEFXN BFFAM BGNUA BKEBE BPEOZ CBEJK CHZPO IEGSK IJVOP OCL RIE RIL RIO |
| ID | FETCH-LOGICAL-i203t-1fdac9fc9d0ceda21dc3e7804fcdda8d37181171ece2536c87aa56bdda967ad43 |
| IEDL.DBID | RIE |
| ISICitedReferencesCount | 1441 |
| ISICitedReferencesURI | http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=000529484005016&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D |
| IngestDate | Wed Sep 10 07:40:26 EDT 2025 |
| IsPeerReviewed | false |
| IsScholarly | true |
| Language | English |
| LinkModel | DirectLink |
| MergedId | FETCHMERGED-LOGICAL-i203t-1fdac9fc9d0ceda21dc3e7804fcdda8d37181171ece2536c87aa56bdda967ad43 |
| PageCount | 10 |
| ParticipantIDs | ieee_primary_8954166 |
| PublicationCentury | 2000 |
| PublicationDate | 2019-June |
| PublicationDateYYYYMMDD | 2019-06-01 |
| PublicationDate_xml | – month: 06 year: 2019 text: 2019-June |
| PublicationDecade | 2010 |
| PublicationTitle | Proceedings (IEEE Computer Society Conference on Computer Vision and Pattern Recognition. Online) |
| PublicationTitleAbbrev | CVPR |
| PublicationYear | 2019 |
| Publisher | IEEE |
| Publisher_xml | – name: IEEE |
| SSID | ssj0003211698 |
| Score | 2.6401565 |
| Snippet | Cascade is a classic yet powerful architecture that has boosted performance on various tasks. However, how to introduce cascade to instance segmentation... |
| SourceID | ieee |
| SourceType | Publisher |
| StartPage | 4969 |
| SubjectTerms | Categorization Computer architecture Computer vision Convolutional codes Instance segmentation Object detection Pattern recognition Recognition: Detection Retrieval Semantic segmentation |
| Title | Hybrid Task Cascade for Instance Segmentation |
| URI | https://ieeexplore.ieee.org/document/8954166 |
| WOSCitedRecordID | wos000529484005016&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D |
| hasFullText | 1 |
| inHoldings | 1 |
| isFullTextHit | |
| isPrint | |
| link | http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV1LSwMxEB7a4sFT1VZ8k4NH1242m9e5WCpIKVpLbyWbzEoRt6XbCv57k-1SPXjxFhJImIRkHvm-GYBbF95FjgHd77yDksQ8MkrLyNDUZRk1mFdo9-mTHI3UbKbHDbjbc2EQsQKf4X1oVn_5bmm3IVTWU5p7-0E0oSml2HG19vEU5j0ZoVWdvYfGutefjp8DdiskpOShQtCv8imV9hi0_7fuEXR_aHhkvFcwx9DA4gTatd1I6ltZdiAafgXeFZmY8p30TRkg78Qbo-Sxsv38LC_49lGTjIouvA4eJv1hVJdBiBZJzDYRzZ2xOrfaxRadSaizDEPeoNw6Z5RjXr1QKilaTDgTVkljuMj8mBbSuJSdQqtYFngGhCmnXG4pQ2pSiWmmGU-NpYjeH82YOIdOkH6-2mW6mNeCX_zdfQmHYXt3wKkraG3WW7yGA_u5WZTrm-p4vgF0QJHW |
| linkProvider | IEEE |
| linkToHtml | http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV3fS8MwED7mFPRp6ib-tg8-Wtc0TZs8D2XDOYbOsbeR5q4yxE7WTfC_N-nK9MEX30ICFy4hubvk--4ArtHdi4Icuh9tgBIGwtdSJb5mEaYp05SVaPdxPxkM5GSihjW42XBhiKgEn9Gta5Z_-Tg3K_dU1pZKWP8h3oJtEVmxa7bW5kWF21gmVrLK38MC1e6Mh08OveVSUgpXI-hXAZXSftw3_jfzPrR-iHjecGNiDqBG-SE0Ks_Rq85l0QS_--WYV95IF29eRxcO9O5Zd9Trld6flfJMr-8VzShvwcv93ajT9atCCP4sDPjSZxlqozKjMDCEOmRoOLnMQZlB1BK5NTCMJYwMhYLHRiZaizi1YypONEb8COr5PKdj8LhEiZlhnJiOEopSxUWkDSOyEWnK4xNoOu2nH-tcF9NK8dO_u69gtzt67E_7vcHDGey5pV7DqM6hvlys6AJ2zOdyViwuy636BjsZlR0 |
| openUrl | ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&rfr_id=info%3Asid%2Fsummon.serialssolutions.com&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Abook&rft.genre=proceeding&rft.title=Proceedings+%28IEEE+Computer+Society+Conference+on+Computer+Vision+and+Pattern+Recognition.+Online%29&rft.atitle=Hybrid+Task+Cascade+for+Instance+Segmentation&rft.au=Chen%2C+Kai&rft.au=Pang%2C+Jiangmiao&rft.au=Wang%2C+Jiaqi&rft.au=Xiong%2C+Yu&rft.date=2019-06-01&rft.pub=IEEE&rft.eissn=1063-6919&rft.spage=4969&rft.epage=4978&rft_id=info:doi/10.1109%2FCVPR.2019.00511&rft.externalDocID=8954166 |