The Design And Implementation Of A Facial Feature Location System Based On Color

Face detection is a key technology of face information processing field, which iswidely used in the public security department, office system, video, and practicalapplication photos. Since1964, which year the face detection technology has developed，there is a variety of face detection technology. The system is mainly about face areadetection based on skin color, and marking facial features area by the histogram statistics.The system uses the color model and the method based on skin color to realize theface detection. In the YCbCr color space, because the influence of the space to receivebrightness is very small, the skin can produce relatively good clustering to make adistinction between the skin and non-skin color easily. Through constructing color Gaussclustering model， the color similarity calculation can be got. By searching thresholdsegmentation, the binaryzation image can be got. According to the distribution, mark theface region through the histogram statistics.After the face detection, got the facial information, such as the location of eyes, noseand mouth. The accuracy of facial feature location can be improved by using a method offace edge detection. According the differences of facial features color, the organs can beisolated by using statistical analysis method to isolate. Because of the influence ofillumination and noise, the results the system got are very difficult to maintain stability. Sothrough a priori knowledge, use the deformation template, histogram statistics andscreening candidate region to get eyes of the relatively accurate contour. After determiningthe position of the eyes, through searching the color and limiting the area，determine theposition of the nose and mouth. And then the face is identified.