【摘要】 中医学认为"光明润泽者,气也,光泽属神气。健康状况良好的人面部有光泽,面部无光泽的人健康状况较差。传统的中医面部判别都是基于个人的主观评价,缺乏客观数据的支持。文中结合计算机模式识别,图像处理技术辅助面部光泽分析,将人脸面部分成五个区域,包括前额,左脸颊,右脸颊,包括鼻梁的中间区域,下颌区域。用PCA,2DPCA特征提取方法在HSV、RGB、Lab、灰度空间进行实验,采用基于最小距离(KNN),贝叶斯分类器分别对五个区域进行光泽分析。采用投票融合的方法,将每个区域分析得到的光泽结果融合作为最终人脸光泽分析结果。更多还原

【Abstract】 Traditional Chinese Medical( TCM) argue that facial gloss is recognized as the Qi. Facial gloss is correlated with body’s healthy condition. The facial glossy can indicate the internal organs’ situation. A heath people’s face looks glossy and a weakness man’s face looks dull. However,because of lacking of objective data,the traditional way of observing facial gloss mainly depends on clinicians subjective appraises. This situation restricts TCM’s development. This paper divides the face into five parts,forehead,left cheek,right cheek,jaw and middle area. It uses computer vision skills and feature extraction methods like PCA,2DPCA under 4 color space using KNN and Bayesian classification to judge glossy or not respectively,Then it undertakes a fusion approach as the final result.更多还原