Face detection system using 1-dimensional HAAR wavelet transform coeffitionts

Abstract

A face detection system is an indispensible component of any automatic systems that deal with human faces. For example, a face recognition system has to rely on the face detection system to process the input image and locate the area of the image corresponding to a face. This area then becomes the input of the face recognition system for further processing. This paper presents a face detection system using I dimensional Haar wavelet transform coefficients as disciminating feature. A set of sample face images is used to generate the reference Haar wavelet coeficients. This reference is then used to find face candidates in the input image. Our results using this algorithm show that detection rate can reach 84.6%.