基于人脸局部特征和SVM的表情识别

A Face Expression Recognition Method Based on Face Features and Support Vector Machine

提出了一种基于人脸局部特征的表情识别方法.首先选取人脸重要的局部特征,对得到的局部特征进行主成分分析,然后用支持向量机(SVM)设计局部特征分类器来确定测试表情图像中局部特征,同时设计支持向量机(SVM)表情分类器,确定表情图像的所属类别.对JAFFE人脸图像数据库进行仿真实验.结果表明,该方法要优于一般的基于整体特征的人脸表情识别方法.

A method for face expression recognition based on face component features was proposed. Face regional component feature was first selected. The principle component analysis （PCA） coefficients were extracted as feature vectors from the face component image. Then, support vector machine （SVM） was used to design feature classification machine for distinguishing the component regions in the face image. Meanwhile, face expression classification machine was also designed to determine which person the image belongs to. Some experiments have been made on the JAFFE face expression image database. The result showed that this method was better than other methods that used the whole features.