Face Detection Based On Skin Segmentation In Color Images And The Feature Location

Recently, human face detection, as a key technology in human face information processing, is becoming a big problem that gathers more attention in the field of pattern recognition and computer vision. And now it has been applied into other application such as entrance security, video coding, video surveillance and tracking and content based image retrieval etc. The study on face detection in color images is becoming an active research subject with the development of computer technology in recent years. Compared with gray scale images, color images sequence provides more information. However, it should be more robust to different illumination conditions, complex background, face occlusion, expression change and so on. The study on face detection in color images is still a challenging task.In this paper, a human face detection method which is based on skin segmentation and the feature location method are studied in color images. Firstly, have improvement in the algorithm of the lighting compensation, in order to alleviating the effects of lighting. Secondly, by analyzing and comparing the cluster of skin in different color space, a mixed skin model is proposed. After analyzing the connected component and using the filter based on mathematical morphology, the candidate facial regions can be obtained. In the selective a validated stage, we mainly utilize fracas'geometrical characters to roughly filter in these candidate regions. Finally, we further verify these regions by detecting eyes and output the result. At the same time the feature localization is researched and studied. In localizing eyes, approach on eyes similarity and the eyes symmetry is presented .In localizing mouth, considering mouth's chromatic character, so mouth is localized by picking up the chromatic.In the last part of the thesis, we implement the algorithm using Visual C++, and experiment in our face testing set. Experiment results show that the algorithm can detect face with complex background and it has high probability of correct detection.