Face Tracking Based on Fusion Skin Color Model and Optical Flow Algorithm

Abstract

A fusion algorithm based on skin color model and optical flow is proposed to detect and track faces. It uses skin color model to detect faces and adopts optical flow algorithm to estimate the continuity of the video frames and obtain the position of faces in the frames. Taking advantages of the algorithms, the proposed method, in some degrees, gives a solution to the effects of face rotation, partial occlusion, expression variation and illumination on face detecting and face tracking.