Class Similarity Weighted Knowledge Distillation for Continual Semantic Segmentation

Deep learning models are known to suffer from the problem of catastrophic forgetting when they incrementally learn new classes. Continual learning for semantic segmentation (CSS) is an emerging field in computer vision. We identify a problem in CSS: A model tends to be confused between old and new c...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Proceedings (IEEE Computer Society Conference on Computer Vision and Pattern Recognition. Online) S. 16845 - 16854
Hauptverfasser: Phan, Minh Hieu, Ta, The-Anh, Phung, Son Lam, Tran-Thanh, Long, Bouzerdoum, Abdesselam
Format: Tagungsbericht
Sprache:Englisch
Veröffentlicht: IEEE 01.06.2022
Schlagworte:
Computational modeling
Computer vision
Deep learning
grouping and shape analysis; Vision applications and systems
Machine vision
Representation learning
Scene analysis and understanding; Computer vision theory; Deep learning architectures and techniques; Efficient learning and inferences; Representation learning; Segmentation
Semantics
Shape
ISSN:1063-6919
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Abstract Deep learning models are known to suffer from the problem of catastrophic forgetting when they incrementally learn new classes. Continual learning for semantic segmentation (CSS) is an emerging field in computer vision. We identify a problem in CSS: A model tends to be confused between old and new classes that are visually similar, which makes it forget the old ones. To address this gap, we propose REMINDER - a new CSS framework and a novel class similarity knowledge distillation (CSW-KD) method. Our CSW-KD method distills the knowledge of a previous model on old classes that are similar to the new one. This provides two main benefits: (i) selectively revising old classes that are more likely to be forgotten, and (ii) better learning new classes by relating them with the previously seen classes. Extensive experiments on Pascal-Voc 2012 and ADE20k datasets show that our approach outperforms state-of-the-art methods on standard CSS settings by up to 7.07% and 8.49%, respectively.
AbstractList Deep learning models are known to suffer from the problem of catastrophic forgetting when they incrementally learn new classes. Continual learning for semantic segmentation (CSS) is an emerging field in computer vision. We identify a problem in CSS: A model tends to be confused between old and new classes that are visually similar, which makes it forget the old ones. To address this gap, we propose REMINDER - a new CSS framework and a novel class similarity knowledge distillation (CSW-KD) method. Our CSW-KD method distills the knowledge of a previous model on old classes that are similar to the new one. This provides two main benefits: (i) selectively revising old classes that are more likely to be forgotten, and (ii) better learning new classes by relating them with the previously seen classes. Extensive experiments on Pascal-Voc 2012 and ADE20k datasets show that our approach outperforms state-of-the-art methods on standard CSS settings by up to 7.07% and 8.49%, respectively.
Author Ta, The-Anh
Phung, Son Lam
Bouzerdoum, Abdesselam
Tran-Thanh, Long
Phan, Minh Hieu
Author_xml – sequence: 1
  givenname: Minh Hieu
  surname: Phan
  fullname: Phan, Minh Hieu
  email: vmhp806@uowmail.edu.au
  organization: University of Wollongong
– sequence: 2
  givenname: The-Anh
  surname: Ta
  fullname: Ta, The-Anh
  email: anhtt71@fsoft.com.vn
  organization: FPT Software, AIC
– sequence: 3
  givenname: Son Lam
  surname: Phung
  fullname: Phung, Son Lam
  email: phung@uow.edu.au
  organization: University of Wollongong
– sequence: 4
  givenname: Long
  surname: Tran-Thanh
  fullname: Tran-Thanh, Long
  email: long.tran-thanh@warwick.ac.uk
  organization: University of Warwick
– sequence: 5
  givenname: Abdesselam
  surname: Bouzerdoum
  fullname: Bouzerdoum, Abdesselam
  email: a.bouzerdoum@uow.edu.au
  organization: University of Wollongong
BookMark eNotjF1LwzAYRqMouM39Ar3IH2jNR_M2uZT6iQPFTb0cSfu2RtJU2ojs3zvUq3PgPDxzchSHiIScc5ZzzsxF9fr0rARonQsmRM44SDggcw6gCjAFyEMy4wxkBoabE7Kcpg_GmBScg9EzsqmCnSa69r0PdvRpR9_Qd-8JG_oQh--ATYf0yk_Jh2CTHyJth5FWQ0w-ftlA19jbvdd76XqM6XdzSo5bGyZc_nNBXm6uN9Vdtnq8va8uV5kXTKYMVa3bEktdKKENyMIKKYQT1pkWtWrUvjWN4g60QsUR2tqAcE5x7pyuG7kgZ3-_HhG3n6Pv7bjbGl0aoUv5AxtwVD4
CODEN IEEPAD
ContentType Conference Proceeding
DBID 6IE
6IH
CBEJK
RIE
RIO
DOI 10.1109/CVPR52688.2022.01636
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 1665469463
9781665469463
EISSN 1063-6919
EndPage 16854
ExternalDocumentID 9879287
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-e5c8f7e7845289634a2322b2ab9fe85d57e7dd51b685e51e6fc962bb511bb8cd3
IEDL.DBID RIE
ISICitedReferencesCount 46
ISICitedReferencesURI http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=000870783002064&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
IngestDate Wed Aug 27 02:15:11 EDT 2025
IsPeerReviewed false
IsScholarly true
Language English
LinkModel DirectLink
MergedId FETCHMERGED-LOGICAL-i203t-e5c8f7e7845289634a2322b2ab9fe85d57e7dd51b685e51e6fc962bb511bb8cd3
PageCount 10
ParticipantIDs ieee_primary_9879287
PublicationCentury 2000
PublicationDate 2022-June
PublicationDateYYYYMMDD 2022-06-01
PublicationDate_xml – month: 06
  year: 2022
  text: 2022-June
PublicationDecade 2020
PublicationTitle Proceedings (IEEE Computer Society Conference on Computer Vision and Pattern Recognition. Online)
PublicationTitleAbbrev CVPR
PublicationYear 2022
Publisher IEEE
Publisher_xml – name: IEEE
SSID ssj0003211698
Score 2.482424
Snippet Deep learning models are known to suffer from the problem of catastrophic forgetting when they incrementally learn new classes. Continual learning for semantic...
SourceID ieee
SourceType Publisher
StartPage 16845
SubjectTerms Computational modeling
Computer vision
Deep learning
grouping and shape analysis; Vision applications and systems
Machine vision
Representation learning
Scene analysis and understanding; Computer vision theory; Deep learning architectures and techniques; Efficient learning and inferences; Representation learning; Segmentation
Semantics
Shape
Title Class Similarity Weighted Knowledge Distillation for Continual Semantic Segmentation
URI https://ieeexplore.ieee.org/document/9879287
WOSCitedRecordID wos000870783002064&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/eLvHCXMwlV09T8MwED21FQMTHy3iWx4YSds4dmzPhQoJVFW0lG5V7VxQJJqifvD7OSehMLCwWYmlSHey_N7l3juAG6kEOiFkYLi1gRBWB5qjCZxxc25SzV1Ryp48qcFAT6dmWIPbnRYGEYvmM2z7ZfEvP1m6rS-VdYgfG0L4dagrpUqt1q6eEhGTiY2u1HFh13R6k-GzNzPxDVyctwnbeB_mXzNUiiukf_C_jx9C60eLx4a7W-YIapgfw0EFHll1NNdNGBfjLdkoW2REVglbs9ei6kmbHr_rZuzOn-j3sv2NEVxl3pwq866kbIQLCnLmaPG2qARJeQte-vfj3kNQjUwIMt6NNgFKp1OFSgtJTCqOxJwQE7d8bk2KWiaS3iWJDG2sJcoQ49SZmLJEsMta7ZLoBBr5MsdTYDZ0jhMhROUiETmpaT-hK2mNxm7C5Rk0fZBmH6UrxqyKz_nfjy9g32ehbLK6hMZmtcUr2HOfm2y9ui5S-QW8tqDQ
linkProvider IEEE
linkToHtml http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV1NT8JAEJ0gmugJFYzf7sGjBbrdbXfPKMGAhAgiN8Jup6aJgOHD3-9sqejBi7dNu0mTmWz2vem8NwC3MhJohZCe5sZ4QhjlKY7as9pOuE4Ut1kpe9iJul01GuleAe62WhhEzJrPsOqW2b_8eG7XrlRWI36sCeHvwK4Ugvsbtda2ohIQlwm1yvVxfl3XGsPes7MzcS1cnFcJ3Tgn5l9TVLJLpFn63-cPofKjxmO97T1zBAWcHUMph48sP5zLMgyyAZesn05ToquErtlrVvekTe3vyhm7d2f6fdMAxwiwMmdPlTpfUtbHKYU5tbR4m-aSpFkFXpoPg0bLy4cmeCmvBysPpVVJhJESkrhUGIgJYSZu-MToBJWMJb2LY-mbUEmUPoaJ1SHliYCXMcrGwQkUZ_MZngIzvrWcKCFGNhCBlYr2E76SRiusx1yeQdkFafyx8cUY5_E5__vxDey3Bk-dceex276AA5eRTcvVJRRXizVewZ79XKXLxXWW1i9H1qQX
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%3Ajournal&rft.genre=proceeding&rft.title=Proceedings+%28IEEE+Computer+Society+Conference+on+Computer+Vision+and+Pattern+Recognition.+Online%29&rft.atitle=Class+Similarity+Weighted+Knowledge+Distillation+for+Continual+Semantic+Segmentation&rft.au=Phan%2C+Minh+Hieu&rft.au=Ta%2C+The-Anh&rft.au=Phung%2C+Son+Lam&rft.au=Tran-Thanh%2C+Long&rft.date=2022-06-01&rft.pub=IEEE&rft.eissn=1063-6919&rft.spage=16845&rft.epage=16854&rft_id=info:doi/10.1109%2FCVPR52688.2022.01636&rft.externalDocID=9879287