Share Email Print

Proceedings Paper

Block error correction codes for face recognition
Author(s): Wafaa R. Hussein; Harin Sellahewa; Sabah A. Jassim
Format Member Price Non-Member Price
PDF $17.00 $21.00

Paper Abstract

Face recognition is one of the most desirable biometric-based authentication schemes to control access to sensitive information/locations and as a proof of identity to claim entitlement to services. The aim of this paper is to develop block-based mechanisms, to reduce recognition errors that result from varying illumination conditions with emphasis on using error correction codes. We investigate the modelling of error patterns in different parts/blocks of face images as a result of differences in illumination conditions, and we use appropriate error correction codes to deal with the corresponding distortion. We test the performance of our proposed schemes using the Extended Yale-B Face Database, which consists of face images belonging to 5 illumination subsets depending on the direction of light source from the camera. In our experiments each image is divided into three horizontal regions as follows: region1, three rows above the eyebrows, eyebrows and eyes; region2, nose region and region3, mouth and chin region. By estimating statistical parameters for errors in each region we select suitable BCH error correction codes that yield improved recognition accuracy for that particular region in comparison to applying error correction codes to the entire image. Discrete Wavelet Transform (DWT) to a depth of 3 is used for face feature extraction, followed by global/local binarization of coefficients in each subbands. We shall demonstrate that the use of BCH improves separation of the distribution of Hamming distances of client-client samples from the distribution of Hamming distances of imposter-client samples.

Paper Details

Date Published: 26 May 2011
PDF: 11 pages
Proc. SPIE 8063, Mobile Multimedia/Image Processing, Security, and Applications 2011, 80630H (26 May 2011); doi: 10.1117/12.883846
Show Author Affiliations
Wafaa R. Hussein, Univ. of Buckingham (United Kingdom)
Harin Sellahewa, Univ. of Buckingham (United Kingdom)
Sabah A. Jassim, Univ. of Buckingham (United Kingdom)

Published in SPIE Proceedings Vol. 8063:
Mobile Multimedia/Image Processing, Security, and Applications 2011
Sos S. Agaian; Sabah A. Jassim; Yingzi Du, Editor(s)

© SPIE. Terms of Use
Back to Top