Fractal model based face recognition for ubiquitous environments

Abstract

In Ubiquitous environments, it is required to identify the user in order to utilize the knowledge of the user's behavior and situation. In this paper, we propose a novel method of face recognition using dominant facial region extraction and fractal model. In order to improve the performance of the face recognition system, we propose an algorithm to extract the dominant facial region from the face images that includes the most discriminated part of the face, and for each dominant facial region it was presented by its fractal model and stored in database. The fractal model of the dominant facial region is then utilized as fractal facial features for face recognition. To further improve the performance of the face recognition system , we also propose the techniques of weighting mask and DC_Free MSE. Finally, some experimental results are presented and demonstrate the excellent performance of our face recognition approach. © 2008 Springer-Verlag Berlin Heidelberg.