Gender classification in low-resolution surveillance video: In-depth comparison of random forests and SVMs

Christopher D. Geelen, Rob G.J. Wijnhoven, Gijs Dubbelman, Peter H.N. de With

Research output: Chapter in Book/Report/Conference proceedingConference contributionAcademicpeer-review

7 Citations (Scopus)

Abstract

This research considers gender classification in surveillance environments, typically involving low-resolution images and a large amount of viewpoint variations and occlusions. Gender classification is inherently difficult due to the large intraclass variation and interclass correlation. We have developed a gender classification system, which is successfully evaluated on two novel datasets, which realistically consider the above conditions, typical for surveillance. The system reaches a mean accuracy of up to 90% and approaches our human baseline of 92.6%, proving a high-quality gender classification system. We also present an in-depth discussion of the fundamental differences between SVM and RF classifiers. We conclude that balancing the degree of randomization in any classifier is required for the highest classification accuracy. For our problem, an RF-SVM hybrid classifier exploiting the combination of HSV and LBP features results in the highest classification accuracy of 89.9±0.2%, while classification computation time is negligible compared to the detection time of pedestrians.

Original languageEnglish
Title of host publicationVideo Surveillance and Transportation Imaging Applications 2015
EditorsRobert P. Loce, Eli Saber
PublisherSPIE
Number of pages15
ISBN (Electronic)9781628414974
DOIs
Publication statusPublished - 1 Jan 2015
EventVideo Surveillance and Transportation Imaging Applications 2015 - San Francisco, United States
Duration: 10 Feb 201512 Feb 2015

Publication series

NameProceedings of SPIE
Volume9407

Conference

ConferenceVideo Surveillance and Transportation Imaging Applications 2015
Country/TerritoryUnited States
CitySan Francisco
Period10/02/1512/02/15

Fingerprint

Dive into the research topics of 'Gender classification in low-resolution surveillance video: In-depth comparison of random forests and SVMs'. Together they form a unique fingerprint.

Cite this