Frontal Face Recognition from Video

TitleFrontal Face Recognition from Video
Publication TypeJournal Article
Year of Publication2008
AuthorsMajumdar, A., and P. Nasiopoulos
Secondary AuthorsBebis, G., R. Boyle, B. Parvin, D. Koracin, P. Remagnino, F. Porikli, J. Peters, J. Klosowski, L. Arns, Y. K. Chun, T. M. Rhyne, and L. Monroe
JournalAdvances in Visual Computing, PT II, Proceedings
Volume5359
Pagination297–306
ISSN0302-9743
Abstract

This work aims at frontal face recognition from video. We propose a new Image-to-Image based recognition approach which is both fast and accurate. We use color information for face recognition. Our feature extraction scheme is robust to changes in absolute color values because it uses curvelet transform coefficients to provide edge based representation. This representation makes our scheme invariant to changes in illumination or tanning. Classification is performed by Kernel classifiers such as the Support Vector Machines and a newly proposed Random classifier. As a result, our scheme eliminates the time consuming dimensionality-reduction step (widely used in face recognition), since it is independent of the dimensionality of the input features. Moreover, our parallel architecture allows for computational benefits as well as the ability to integrate depth information in the future. A very short sequence of video (2 seconds) is required for face authentication. Performance evaluations using a standard frontal-face video database show a recognition accuracy of around 99.9%. For short frontal-face video sequences, the proposed scheme outperforms current video based recognition systems by 20%.

a place of mind, The University of British Columbia

Electrical and Computer Engineering
2332 Main Mall
Vancouver, BC Canada V6T 1Z4
Tel +1.604.822.2872
Fax +1.604.822.5949
Email:

Emergency Procedures | Accessibility | Contact UBC | © Copyright 2021 The University of British Columbia