A face recognition system : a fourier-DCT and wavelet-walsh based approaches

Abstract

In our daily life, face recognition systems become very important. It is a biometric method to extract the main features of the faces and identify the individuals. Many algorithms were developed in order to have a high accuracy of identification. This project describes two new proposed methods. The first approach is based on Discrete Fourier Transform, Polar mapping and Discrete Cosine Transform or Walsh Hadamard Transform. The second approach requires the implementation of a Wavelet transform followed by the Walsh Transform. The algorithms are evaluated using faces images from a well-known database. The results are also compared with previous developed approaches. All the algorithms are implemented using MATLAB. The comparison shows a better performance for the proposed approaches over the previous techniques.