An Architecture for Real Time Face Recognition Using WMPCA

An architecture for real time face recognition using weighted modular principle component analysis (WMPCA) is presented in this paper. The WMPCA methodology splits the test face horizontally into sub-regions and analyzes each sub-region separately using PCA. The final decision is taken based on a weighted sum of the errors obtained from each region. This is based on assumption that different regions in a face vary at different rates with variations in expression and illumination. The WMPCA methodology has a better recognition rate, when compared with conventional PCA, for faces with large variations in expression and illumination. This methodology has a wide scope for parallelism. An architecture which exploits this parallelism is proposed in this paper. We also present a System On Programmable Chip (SOPC) implementation of face recognition system using this architecture.