Face Detection Based on Self-Skin Segmentation and Wavelet Support Vector Machine

This paper addresses the problem of face detection in color image. In this paper, a coarse-to-fine approach to face detection is presented. Firstly, a novel self-skin segmentation algorithm is proposed for skin detection, instead of conventional skin color model. In the second step, edge information and mathematical morphology method are integrated to progressively restrict the possible candidate face regions to smaller areas. Finally, support vector machine based on a wavelet basis is used for face verification. Face images for test are classified as simple, mid-complicated and complicated images. Experiments show that this algorithm presented in this paper is of accuracy and reliability