Research On System Of Face Detection And Tracking

Face detection and face tracking is a foundational and key technology in the facial analysis,thus can be applied in many fields,such as video surveillance,human computer interraction,biometrics,content based image retrieval,image coding,etc.In this paper ,a deeply research on face detection in color images and face tracking in video sequences is proposed , the main work comprises the following two parts:1. The typical algorithms of face detection are given,with the emphasis on color-based face detection algorithm and AdaBoost face detection algorithm . For face detection in static image, the color model for verification is used to reduce the rate of error detection in AdaBoost face detection algorithm. For face detection in vedio sequences, to improve the speed of face detection,the skin color detection is used for quick locating,then the Haar-like features and AdaBoost classifier are used to detect faces.2. For face tracking in video sequences, the Mean Shift tracking algorithm is investigated in detail.An improved CamShift algorithm based on Mean-Shift is introduced ,and then using color histogram as tracking clues, thus get rapid and accurate face tracking. Finally on the application environment based on VC + + 6.0 ,the experimental system of the automatic single-face detection and tracking is developed.The experimental system used a Samsung camera and the OK series of PCI graphics card to capture a real-time video image and the track-ing speed is over 15 frames per-second.