Research Of Face Detection In Video Moving Pictures

In this thesis, a study on the algorithms to face detection is presented. Face detection has been a research focus of computer vision, pattern recognition and human-computer interaction for a long time. It also has the application in a wide range of fields such as automatic face recognition, visual surveillance and content-based retrieval. However, face detection is a most challenging task because of the complicated pattern and the frangibility of human face. So, most face detection methods have the weakness of large computation, low efficiency and many false reports among the detection result.Based on the survey of the existing methods, a study on the on the application of skin-color model to face detection is proposed. The application of the skin-color information has the remarkable advantage of efficiency So it can accelerate the detection and make the false report decrease without cutting down the accuracy.The focus of this thesis is the application of motion vectors of MPEG stream and skin-color model to the pretreatment of the face detection A detailed discussion on related theories of digital image processing is presented. Then a fast and robust face detection algorithm is proposed At last, a prototype application named with "Golden Eyes Face Detection System" is designed with VC6.0 to verify the validity and accuracy of the algorithm.At the same time, major functions of Digital Video Record(DVR) are implemented in the system too.