作者: Zhong Jin , Zhen Lou , Jingyu Yang , Quansen Sun
DOI: 10.1016/J.NEUCOM.2006.10.043
关键词: Computer vision 、 Face (geometry) 、 Face detection 、 Color histogram 、 FERET database 、 Pixel 、 Rectangle 、 Mathematics 、 Pattern recognition 、 Artificial intelligence 、 Color image 、 Template matching
摘要: A face-detection approach is proposed in the paper. Firstly, a luminance-conditional distribution model of skin-color information used to detect skin pixels color images; then, morphological operations are extract skin-region rectangles; finally, template matching based on linear transformation face each rectangle. Experimental results some images from FERET database and Internet encouraging. The algorithm shown be effective efficient detecting frontal faces images.