Share Email Print

Proceedings Paper

Secure wavelet-based isometric projection for face recognition
Format Member Price Non-Member Price
PDF $17.00 $21.00

Paper Abstract

Biometric systems such as face recognition must address four key challenges: efficiency, robustness, accuracy and security. Isometric projection has been proposed as a robust dimension reduction technique for a number of applications, but it is computationally demanding when applied to high dimensional spaces such as the space of face images. On the other hand, wavelet transforms have shown to provide an efficient tool for facial feature representation and face recognition with significant reduction in dimension. In this paper, we propose a hybrid approach that combines the efficiency and robustness of wavelet transforms with isometric projections for face features extraction in the transformed domain to be used for recognition. We shall compare the recognition accuracy of our approach with the accuracy of other commonly used projection techniques in the wavelet domain such as PCA and LDA. The security of biometric templates is addressed by adopting a lightweight random projection technique as an add-on subsystem. The results are based on experiments conducted on a publicly available benchmark face database.

Paper Details

Date Published: 31 May 2011
PDF: 9 pages
Proc. SPIE 8063, Mobile Multimedia/Image Processing, Security, and Applications 2011, 80630U (31 May 2011); doi: 10.1117/12.884576
Show Author Affiliations
Hisham Al-Assam, 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
Sign in to read the full article
Create a free SPIE account to get access to
premium articles and original research
Forgot your username?