面诊是中医四诊之一,在中医面诊时通常将人脸区域分成5个部分,每个部分对应于不同的人体内脏部位,内脏的病变表现为人脸对应区域的颜色变化.实现了一种简单快速的人脸区域分割方法,首先用Gabor小波变换定位人眼,将AdaBoost和唇色模型相结合定位嘴角,然后根据先验知识确定内眉点位置,用提取轮廓算法确定人脸边界,从而实现人脸区域的自动分割.通过对样本图片的测试,实验结果表明该方法可以满足面诊实时分割人脸区域的需求.%Face diagnosis is one of the four diagnostic methods for traditional Chinese medicine (TCM). Face region is usually divided into five parts during face diagnosis process, each part corresponding to the different human organs. The morbidity of the organs can be represented from the color of related face region. A simple and fast face region segmentation method is proposed. First, Gabor wavelet transform is used to locate eye. Ada-Boost is combined with lip color model to locate mouth corner. The eyebrow coordinate is confirmed based on the priori knowledge. Then the face is segmented automatically after the face boundary is determined using the contour extraction algorithm. The experimental results show that the proposed method can satisfy the real-time segmentation requirements of face diagnosis.
展开▼
