This paper presents an automatic approach to detect face facial features from face images using the skin color information and active contour model. Skin color has been widely used for face detection since it provides an important clue for face detection and needs less computational burden. However, the variations of lighting condition cause the difficulties in robust detection of skin colors. Meanwhile, active contour model is regarded as an efficient way to detect object boundaries even though the final result is largely depends on an initial boundary condition. In this paper, we propose a robust face detection technique by compensating for varying light conditions and utilizing the normalized YCbCr color model to detect candidate regions of face and facial feature. In order to extract optimal boundaries efface and facial feature, we adopt active contour model. The initialization of the active contour is done without any i nteraction and the m inimal face and facial feature region assigned as the initial set of active contour resolves the initial set problem of the model. The experimental results show the efficiency of the proposed method. The face and facial feature information will be used for face recognition and facial feature descriptor.
展开▼
