Face Localization Using Backpropagation

Abstract

There is more than one type of neural net uses to classify complex pattern (face), and it is represent very suitable technique for face localization. We present backpropagation neural network-based face localization system. Face localization in image is a two class (face, nonface) and it is represent very an important first step for many systems ranging from supervision to human computer interface because of the computational model of face. The proposed system consist of two subsystem ,first subsystem is preprocessing operations on image to make it smooth and make next operation (training patterns in backpropagation neural net )easy as possible .Second subsystem is classifying the patterns produced by the preprocessing stage consists showing the patterns(an image) to backpropagation neural net to localize the face in an image .