Face detection process is easily affected by facial details,shape,expression,posture,etc,and the skin color feature,as one of the most recognizable information in human face,is relatively stable.This paper implements face detection and labeling based on the skin color model.The skin color model is con-structed in the YCbCr color space,and the influence of environmental factors such as uneven illumination and brightness changes in the process of skin color extraction is eliminated through light compensation,and the noise in the non-face area is filtered based on morphological filtering.Combined with the geometric fea-tures of the face area,the face positioning and box check mark are realized.The test results of face images under various conditions,such as uneven illumination,occlusion,and large angles,show that the imple-mented algorithm can better achieve face detection and localization.
关键词
肤色模型/颜色空间转换/光照补偿/人脸几何特征/人脸检测
Key words
skin color model/color space conversion/light compensation/face geometric features/face detection
维普
万方数据