A Study And Application Of Face Detection On Clothing Show

In recent years,with the development of information technology and applications fields,face detection and tracking has been a research hotspot in computer application domain. Related research results come out continuously.This paper reviews the development of face detection and tracking, being used in the video of clothing show. First, analyzes from the domestic and international research papers on face detection and tracking in recent years in this paper. Then, discussed their advantages and disadvantages. Finally, a face detection and tracking system is developed here. My primary work is as following:1,Study and comparison of current face detection algorithms, this paper proposed a fast and simple face detection method which uses skin model to attain face candidates and uses the Cascade classifier to verify the human face.In this method, a color transformation technique is applied to the face images in the YCbCr chrominance space. Then the skin regions could be segmented by a series of threshold to the skin tone model. On considering to different lighting conditions, we adopt lighting compensation technique solve the problem. After the face candidate areas are obtained, a method to get their sizes quickly is proposed here and then face candidate areas would be signed.2,Using face detection method based on Cascade classifier,which fast calculate features by"integral image"method, then generating simple classifier and optimizing its weight, to build a efficient strong classifier. Eventually, make the individual strong classifier to cascade into a more robust cascade classifier which was used in face detection.Focusing on the problem of long training time in face detection with AdaBoost algorithm, presents a fast AdaBoost training algorithm. Experimental result show that the algorithm gains higher training efficiency.3,Kalman algorithm is used in the paper for face tracking, which combines both the face detection and tracking. Face detection algorithm is used to determine the initial tracking regional: after Kalman algorithm forecasts the face's information in the next frame. Then matches are done to make sure whether the tracking is right or not. If right, the face records should be renewed and the results are signed. Experimental result show Kalman algorithm is simple calculation, fast tracking.