自动人脸表情识别中的面部特征点精确定位与混合表情特征集优化.PDFVIP

2010 First International Conference on Cellular, Molecular Biology, Biophysics and Bioengineering (CMBB) Improve the localization accuracy of facial feature and optimize hybrid expression features for automatic facial expression recognition 1 2* 2 2 Ling Zhang , Siping Chen , Tianfu Wang , and Zhuo Liu 1) Department of Biomedical Engineering, Zhejiang University, Hangzhou, Zhejiang, China (zhangling0722@163.com) 2) Shenzhen Key Lab of Biomedical Engineering, Shenzhen University, Shenzhen, Guangdong, China (chensiping@) Abstract—This paper describes an automatic recognition approach for facial expression from nearly front view image. This work has two contributions. The first is a method for adaptively generating the initial model for Active Appearance Models (AAM) fitting, which allows the facial feature under large variation of facial expression to be detected more accuracy. The second is the introduction of a set of hybrid expression features consist of geometric features, AAM shape and appearance parameters and which can be optimize when adopting quadratic mutual information (QMI) feature selection method. The experimental results on CAS-PEAL facial expression database show that by using support vector machine (SVM) classifier, the proposed method can achieve the recognition rate of 87.33%. Keywords—facial expression recognition; AAM initial model; facial feature localization; hybrid expression feature; quadratic mutual information 自动人脸表情识别中的面部特征点精确定位与混合表 情特征集优化 张灵 1 陈思平 2* 汪天富 2 刘卓 2 1) 浙江大学生物医学工程系，杭州，浙江，中国 2) 深圳大学生物医学工程重点实验室，深圳，广东，中国 摘 要 本文提出了一个针对静态准正面人脸图像的全自动表情识别方法。主要有两个贡献：其一，提出了一个主动表观模型 （AAM ）中自适应生成初始模型的方法