Show simple item record

dc.contributor.authorXiong, Jinhui
dc.contributor.authorRichtarik, Peter
dc.contributor.authorHeidrich, Wolfgang
dc.date.accessioned2019-11-27T12:25:34Z
dc.date.available2019-11-27T12:25:34Z
dc.date.issued2019-09-29
dc.identifier.citationXiong, J., Richtarik, P., & Heidrich, W. (2019). Stochastic Convolutional Sparse Coding. Vision, Modeling and Visualization. https://doi.org/10.2312/VMV.20191317
dc.identifier.isbn978-3-03868-098-7
dc.identifier.doi10.2312/vmv.20191317
dc.identifier.urihttp://hdl.handle.net/10754/660286
dc.description.abstractState-of-the-art methods for Convolutional Sparse Coding usually employ Fourier-domain solvers in order to speed up the convolution operators. However, this approach is not without shortcomings. For example, Fourier-domain representations implicitly assume circular boundary conditions and make it hard to fully exploit the sparsity of the problem as well as the small spatial support of the filters. In this work, we propose a novel stochastic spatial-domain solver, in which a randomized subsampling strategy is introduced during the learning sparse codes. Afterwards, we extend the proposed strategy in conjunction with online learning, scaling the CSC model up to very large sample sizes. In both cases, we show experimentally that the proposed subsampling strategy, with a reasonable selection of the subsampling rate, outperforms the state-of-the-art frequency-domain solvers in terms of execution time without losing the learning quality. Finally, we evaluate the effectiveness of the over-complete dictionary learned from large-scale datasets, which demonstrates an improved sparse representation of the natural images on account of more abundant learned image features.
dc.description.sponsorshipThis work was supported by King Abdullah University of Science and Technology as part of VCC center baseline funding.
dc.publisherThe Eurographics Association
dc.relation.urlhttps://diglib.eg.org/handle/10.2312/vmv20191317
dc.rightsArchived with thanks to arXiv
dc.titleStochastic Convolutional Sparse Coding
dc.typeConference Paper
dc.contributor.departmentComputer Science Program
dc.contributor.departmentComputer, Electrical and Mathematical Sciences and Engineering (CEMSE) Division
dc.contributor.departmentVisual Computing Center (VCC)
dc.conference.dateSep 30 - Oct 2, 2019
dc.conference.nameVision, Modeling and Visualization 2019
dc.conference.locationRostock, Germany
dc.eprint.versionPre-print
dc.identifier.arxivid1909.00145
kaust.personXiong, Jinhui
kaust.personRichtarik, Peter
kaust.personHeidrich, Wolfgang
refterms.dateFOA2019-11-27T12:26:38Z
dc.date.posted2019-08-31


Files in this item

Thumbnail
Name:
Preprintfile1.pdf
Size:
3.077Mb
Format:
PDF
Description:
Pre-print
Thumbnail
Name:
paper_1010_supplement.pdf
Size:
2.176Mb
Format:
PDF
Description:
Supplementary Material

This item appears in the following Collection(s)

Show simple item record