Systolic array realization of a neural network-based face recognition system

Automatic recognition of faces has wide applications ranging from security to man-machine interaction. A popular method for face recognition is using eigen faces. Eigen faces are obtained by doing principal component analysis (PCA) on the face database. A neural network that performs PCA is called principal component neural network (PCNN). This paper presents a systolic array design for principal component neural network-based face recognition system. Results of implementation of the design in an FPGA device of Xilinx confirm the suitability of the design for video surveillance in busy public places.