A New Face Decomposition for Face Detection using SVM Classifier

Face detection is a crucial step in the process of Biometric systems. The aim of this work is to study some methods of feature extraction for face detection using Local Discrete Cosine Transform named LDCT, Global Discrete Cosine Transform named GDCT and Discrete Wavelet Transform called DWT using three different levels (2, 3 and 4), we have used each one separately and we have combined both of them to evaluate the performance of the proposed method, then comparison between these methods will be exposed. Our new proposed method aims to select feature from the extracted ones to reduce dimensionality, we have proceeded to select the significant features in face, like eyes mouth and nose. Finally, we applied Support Vector Machine (SVM) classifier to discriminate between two classes, faces and non-faces. We present experimental results applied on MIT face database to demonstrate the effectiveness of the novel approach in terms of accuracy and running time.