基于特征区域自动分割的人脸表情识别

摘要

针对目前三维人脸表情区域分割方法复杂、费时问题,提出一种人脸表情区域自动分割方法,通过投影、曲率计算的方法检测人脸的部分特征点,以上述特征点为基础进行人脸表情区域的自动分割.为得到更加丰富的表情特征,结合人脸表情识别编码规则对提取到的特征矩阵进行扩充,利用分类器进行人脸表情的识别.通过对三维人脸表情数据库部分样本的识别结果表明,该方法可以取得较高的识别率.%To improve 3D facial expression feature regions segmentation, an automatic feature regions segmentation method is presented.The facial feature points are detected by conducting projection and curvature calculation, and are used as the basis of facial expression feature regions automatic segmentation.To obtain more abundant facial expression information, the Facial Action Coding System(FACS) coding roles is introduced to extend the extracted characteristic matrix.And facial expressions can be recognized by combining classifiers.Experimental results of 3D facial expression samples show that the method is effective with high recognition rate.