A fast face detection method combining skin color feature and AdaBoost
Saiping Ji, Xiaobo Lu, Qianzhou Xu · 2014
Face detection technology is a hot topic of pattern recognition and computer vision. The method based on skin color feature and AdaBoost are both the most representative methods. Among them, the former method has a faster detection speed, but lower accuracy; the latter method has a higher accuracy, but slower speed. In this paper, a fast face detection method based on the combination of skin color feature and AdaBoost is proposed. Skin color feature method is the front level for the AdaBoost method so as to reduce the detection range of AdaBoost method and shorten the time of face detection. On this basis, Gaussian pyramid is also introduced into this method. By means of creating Gaussian pyramid for the image to be detected, skin color feature and AdaBoost methods are performed on different layers of Gaussian pyramid. The experiments were performed on our own face database. The result showed that the proposed method took account of the speed and accuracy and improved the efficiency of face detection.