Research of Face Detection Based on Skin Color Feature

Abstract

Abstract Face image contains many pattern features and this paper studies a face detection algorithm based on skin color features. Through the analysis of the face detection method based on skin color features, the calculation process and principle are understood, and the algorithm and detection techniques such as color space, preprocessing face image detection technology. Building a corresponding skin color model based on face characteristics are analyzed in depth. Because the skin color only occupies a specific range in the primary color of color space, the probability likelihood graph of skin color is obtained by calculating skin color similarity. Then the binary image of the possible face region is obtained by smooth filtering, threshold segmentation, and other operations. Then the area center and height ratio are used to calculate the center of mass and deflection angle, and the correlation coefficient is matched with the face module. Finally, the face area is detected and marked. Test results show that the algorithm can achieve face detection well.