Representation Learning by Learning to Count

Simple item page
dc.contributor.authorNoroozi, Mehdi
dc.contributor.authorPirsiavash, Hamed
dc.contributor.authorFavaro, Paolo
dc.date.accessioned2019-07-01T17:51:38Z
dc.date.available2019-07-01T17:51:38Z
dc.date.issued2017-12-25
dc.description.abstractWe introduce a novel method for representation learning that uses an artificial supervision signal based on counting visual primitives. This supervision signal is obtained from an equivariance relation, which does not require any manual annotation. We relate transformations of images to transformations of the representations. More specifically, we look for the representation that satisfies such relation rather than the transformations that match a given representation. In this paper, we use two image transformations in the context of counting: scaling and tiling. The first transformation exploits the fact that the number of visual primitives should be invariant to scale. The second transformation allows us to equate the total number of visual primitives in each tile to that in the whole image. These two transformations are combined in one constraint and used to train a neural network with a contrastive loss. The proposed task produces representations that perform on par or exceed the state of the art in transfer learning benchmarks.en_US
dc.description.sponsorshipPaolo Favaro acknowledges support from the Swiss National Science Foundation on project 200021 149227. Hamed Pirsiavash acknowledges support from GE Global Research.en_US
dc.description.urihttps://ieeexplore.ieee.org/document/8237890en_US
dc.format.extent9 pagesen_US
dc.genreconference papers and proceedings preprintsen_US
dc.identifierdoi:10.13016/m2qaoo-thod
dc.identifier.citationMehdi Noroozi , et.al, Representation Learning by Learning to Count, 2017 IEEE International Conference on Computer Vision (ICCV), DOI: 10.1109/ICCV.2017.628en_US
dc.identifier.urihttps://doi.org/10.1109/ICCV.2017.628
dc.identifier.urihttp://hdl.handle.net/11603/14327
dc.language.isoen_USen_US
dc.publisherIEEEen_US
dc.relation.isAvailableAtThe University of Maryland, Baltimore County (UMBC)
dc.relation.ispartofUMBC Computer Science and Electrical Engineering Department Collection
dc.relation.ispartofUMBC Faculty Collection
dc.rightsThis item is likely protected under Title 17 of the U.S. Copyright Law. Unless on a Creative Commons license, for uses protected by Copyright Law, contact the copyright holder or the author.
dc.rights© 2017 IEEE
dc.subjectimage representationen_US
dc.subjectneural netsen_US
dc.subjectrepresentation learningen_US
dc.subjectartificial supervision signalen_US
dc.subjectequivariance relationen_US
dc.subjectmanual annotationen_US
dc.subjectvisual primitivesen_US
dc.titleRepresentation Learning by Learning to Counten_US
dc.typeTexten_US

Files

Original bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
1708.06734.pdf
Size:
4.08 MB
Format:
Adobe Portable Document Format
Description:
Download
License bundle
Now showing 1 - 1 of 1
No Thumbnail Available
Name:
license.txt
Size:
2.56 KB
Format:
Item-specific license agreed upon to submission
Description:
Download

Collections

UMBC Computer Science and Electrical Engineering Department
UMBC Faculty Collection