Face detection based on skin color segmentation and AdaBoost algorithm
Guanghui Xu, Yingcai Xiao, Shuai Xie, Sen Zhu · 2017
In practical applications, face detection based on AdaBoost algorithm usually has high false positive rate and missing rate due to the interference of complex background in color images. To address these problems, this paper proposes a face detection method combined skin color segmentation with the AdaBoost algorithm. Firstly, in order to avoid the influence of poor lighting conditions, we use the Reference White algorithm to compensate the illumination of the input image. Then we convert the color space from RGB to YCgCr and segment the image into skin regions and non-skin regions. The AdaBoost algorithm is applied to train classifiers. After the skin color segmentation, the trained classifiers are used to detect faces. The experimental results show that this combined method can effectively reduce the rate of both false and miss detection.