Multi-resolution feature extraction in human face

Abstract

This paper presents a new method for feature extraction. In the first stage, the coarse face region is found from original image using general method. In the second stage, the proposed method finds the coarse eye and mouth region from multiresolution image using wavelet transformation based on the edge distribution information in the different scale image. Then, the eye and mouths' special site in the original image can be obtained. In the third stage, extract features of eye and mouth. The experimental results have shown that the proposed method can locate eyes accurately from the input image with complex background. At the same time, this proposed method is needed not to create the particular wavelet function.