JPEG Compression Effects on a Smart Card Face Verification System

We discuss the effect of a compression scheme to be applied to image data in a smart card based face verification system. Accordingly, we propose a system architecture considering trade-off between performance versus the improvement of memory and bandwidth management. In order to establish the system limitations, studies were performed on BANCA, XM2VTS and FERET databases demonstrating that by using JPEG compression on face representations, the verification performance of the system is not degraded and that the proposed architecture is a viable one allowing also the possibility of non-optimum selection of normalisation parameters in the preprocessing stage. Different results were achieved on different databases, indicating that the use of different operating scenarios for system evaluation may call for different optimum operating points.

[1]  Azriel Rosenfeld,et al.  Face recognition: A literature survey , 2003, CSUR.

[2]  Rama Chellappa,et al.  Biometric image processing and recognition , 1998, 9th European Signal Processing Conference (EUSIPCO 1998).

[3]  Jean-Philippe Thiran,et al.  The BANCA Database and Evaluation Protocol , 2003, AVBPA.

[4]  Josef Kittler,et al.  Face verification system architecture using smart cards , 2004, Proceedings of the 17th International Conference on Pattern Recognition, 2004. ICPR 2004..

[5]  Touradj Ebrahimi,et al.  A study of JPEG 2000 still image coding versus other standards , 2000, 2000 10th European Signal Processing Conference.

[6]  Jiri Matas,et al.  XM2VTSDB: The Extended M2VTS Database , 1999 .

[7]  Josef Kittler,et al.  Performance versus Computational Complexity Trade-Off in Face Verification , 2004, ICBA.