Skin Color Distribution and Template Matching Based Face Detection Method

摘要

Face detection is one of the challenge tasks for computer vision and pattern recognition. Especially in recognition of size unavailable face of the normal scenery. In this paper, we present a skin color distribution and template matching based face detecting algorithm. Firstly, skin color information of the face is used for a superficial detection on original color image to segment regions in which may have face. 'Eyes in whole and face itself template is employed to complete target recognition in effective area after segmentation. It's matching degree standard is the correlation coefficient of face template and window image. Finally, use radial basis inner product function as the Support Vector Machine (SVM) classifier of the sorting algorithm to precisely locate the face. The experimental result shows, this method effectively accelerate the detecting speed, and has higher robustness.