Show simple item record

dc.contributor.authorDubey, Abhimanyu
dc.contributor.authorGupta, Otkrist
dc.contributor.authorGuo, Pei
dc.contributor.authorRaskar, Ramesh
dc.contributor.authorFarrell, Ryan
dc.date.accessioned2021-11-10T15:56:28Z
dc.date.available2021-11-10T12:34:50Z
dc.date.available2021-11-10T15:56:28Z
dc.date.issued2018-07
dc.identifier.urihttps://hdl.handle.net/1721.1/138096.2
dc.description.abstract© Springer Nature Switzerland AG 2018. Fine-Grained Visual Classification (FGVC) datasets contain small sample sizes, along with significant intra-class variation and inter-class similarity. While prior work has addressed intra-class variation using localization and segmentation techniques, inter-class similarity may also affect feature learning and reduce classification performance. In this work, we address this problem using a novel optimization procedure for the end-to-end neural network training on FGVC tasks. Our procedure, called Pairwise Confusion (PC) reduces overfitting by intentionally introducing confusion in the activations. With PC regularization, we obtain state-of-the-art performance on six of the most widely-used FGVC datasets and demonstrate improved localization ability. PC is easy to implement, does not need excessive hyperparameter tuning during training, and does not add significant overhead during test time.en_US
dc.language.isoen
dc.publisherSpringer International Publishingen_US
dc.relation.isversionof10.1007/978-3-030-01258-8_5en_US
dc.rightsCreative Commons Attribution-Noncommercial-Share Alikeen_US
dc.rights.urihttp://creativecommons.org/licenses/by-nc-sa/4.0/en_US
dc.sourcearXiven_US
dc.titlePairwise confusion for fine-grained visual classificationen_US
dc.typeArticleen_US
dc.identifier.citationDubey, Abhimanyu, Gupta, Otkrist, Guo, Pei, Raskar, Ramesh and Farrell, Ryan. 2018. "Pairwise confusion for fine-grained visual classification."en_US
dc.contributor.departmentProgram in Media Arts and Sciences (Massachusetts Institute of Technology)en_US
dc.eprint.versionOriginal manuscripten_US
dc.type.urihttp://purl.org/eprint/type/ConferencePaperen_US
eprint.statushttp://purl.org/eprint/status/NonPeerRevieweden_US
dc.date.updated2019-08-02T14:26:17Z
dspace.date.submission2019-08-02T14:26:18Z
mit.metadata.statusPublication Information Neededen_US


Files in this item

Thumbnail

This item appears in the following Collection(s)

Show simple item record

VersionItemDateSummary

*Selected version