Robust Face Tracking Using Particle Filter Based on Deformable Face Model and Skin Color

This paper presents an algorithm for real-time and robust human face tracking against pictures and other objects. Face detection is based on template matching, morphological operations and skin color segmentation and motion information. Tracing is performed using particle filter dependent upon skin color and PCA model for face. The face detector locates human faces from the face candidates by using motion information. Then the detected face is tracked using particle filter based on skin color and the probabilistic face model, which is updated using the information related to variability with respect to head rotation, illumination, facial expression, and occlusion. The proposed face tracking method is real-time and achieves high performance, robustness to illumination variations and geometric changes (such as viewpoint and scale changes). The experimental results show that the proposed algorithm is the best comparing with other methods specially when there is occlusion. Also the robustness of the suggested method, when the tracked face is close to an equicolor object, is proved