The Technique And Implement Of Face Detection And Tracking In Video

Automatic face detection and tracking has been an important and popular subject in the fields of computer vision and pattern recognition over the past few years. In this thesis, face detection and tracking in video sequences is studied. For the beginning, the background knowledge and classifications of face detection and tracking are given. Then, the face detection arithmetic used in this thesis is introduced in detail. The face tracking in video sequences is divided into two stages: the first one is region segmentation; two ways to segment the regions which may contain faces is studied which are skin-color segmentation and motion region segmentation, and in this thesis motion detection arithmetic is used to segment the regions; the other one is human face detection in the candidate regions using the face detection arithmetic. In the end, a face detection and tracking system is developed. The experimental results show that the system can detect and track the human faces in the video sequences from the camera, so it can be used in video monitoring system.