Supervised Convolutional Sparse Coding

Handle URI:
http://hdl.handle.net/10754/627542
Title:
Supervised Convolutional Sparse Coding
Authors:
Affara, Lama Ahmed; Ghanem, Bernard ( 0000-0002-5534-587X ) ; Wonka, Peter ( 0000-0003-0627-9746 )
Abstract:
Convolutional Sparse Coding (CSC) is a well-established image representation model especially suited for image restoration tasks. In this work, we extend the applicability of this model by proposing a supervised approach to convolutional sparse coding, which aims at learning discriminative dictionaries instead of purely reconstructive ones. We incorporate a supervised regularization term into the traditional unsupervised CSC objective to encourage the final dictionary elements to be discriminative. Experimental results show that using supervised convolutional learning results in two key advantages. First, we learn more semantically relevant filters in the dictionary and second, we achieve improved image reconstruction on unseen data.
KAUST Department:
Computer, Electrical and Mathematical Sciences and Engineering (CEMSE) Division; Computer Science Program; Electrical Engineering Program; Visual Computing Center (VCC)
Publisher:
arXiv
Issue Date:
8-Apr-2018
ARXIV:
arXiv:1804.02678
Type:
Preprint
Additional Links:
http://arxiv.org/abs/1804.02678v1; http://arxiv.org/pdf/1804.02678v1
Appears in Collections:
Other/General Submission; Computer Science Program; Electrical Engineering Program; Visual Computing Center (VCC); Computer, Electrical and Mathematical Sciences and Engineering (CEMSE) Division

Full metadata record

DC FieldValue Language
dc.contributor.authorAffara, Lama Ahmeden
dc.contributor.authorGhanem, Bernarden
dc.contributor.authorWonka, Peteren
dc.date.accessioned2018-04-16T11:27:44Z-
dc.date.available2018-04-16T11:27:44Z-
dc.date.issued2018-04-08en
dc.identifier.urihttp://hdl.handle.net/10754/627542-
dc.description.abstractConvolutional Sparse Coding (CSC) is a well-established image representation model especially suited for image restoration tasks. In this work, we extend the applicability of this model by proposing a supervised approach to convolutional sparse coding, which aims at learning discriminative dictionaries instead of purely reconstructive ones. We incorporate a supervised regularization term into the traditional unsupervised CSC objective to encourage the final dictionary elements to be discriminative. Experimental results show that using supervised convolutional learning results in two key advantages. First, we learn more semantically relevant filters in the dictionary and second, we achieve improved image reconstruction on unseen data.en
dc.publisherarXiven
dc.relation.urlhttp://arxiv.org/abs/1804.02678v1en
dc.relation.urlhttp://arxiv.org/pdf/1804.02678v1en
dc.rightsArchived with thanks to arXiven
dc.titleSupervised Convolutional Sparse Codingen
dc.typePreprinten
dc.contributor.departmentComputer, Electrical and Mathematical Sciences and Engineering (CEMSE) Divisionen
dc.contributor.departmentComputer Science Programen
dc.contributor.departmentElectrical Engineering Programen
dc.contributor.departmentVisual Computing Center (VCC)en
dc.eprint.versionPre-printen
dc.identifier.arxividarXiv:1804.02678en
kaust.authorAffara, Lama Ahmeden
kaust.authorGhanem, Bernarden
kaust.authorWonka, Peteren
All Items in KAUST are protected by copyright, with all rights reserved, unless otherwise indicated.