Facial Feature Detection and Tracking Based on the Video Image

This paper uses Matlab to extract the signal characteristics from video images and track the human face in order to realize the face detection and face tracking. First, corner feature of human face will be extracted via Min Eigen Features function based on Harris-Stephens algorithm. Then it uses vision Cascade Object Detector function to detect the edge of the human face from an image or the video. At last, the KTL (Kanade-Lucas-Tomasi) tracking method based on the feature of human face is introduced in this paper. This method can track the moving target. According to the research, using Harris-Stephens algorithm can accurately extract the feature of the face. The speed of KTL's calculation method is fast enough to satisfy the real time demand, and the stimulation results show that the method is able to track the object effectively. Therefore, the paper based on face detection and recognition can improve the efficiency of public security organs and it is of great reference significance.