基于LBP和SVM决策树的人脸表情识别  

Facial Expression Recognition Based on LBP and SVM Decision Tree

作　　者：李扬[1] 郭海礁[1] 

机构地区：[1]天津现代职业技术学院信息工程系,天津300350

出　　处：《现代计算机》2014年第6期40-44,共5页Modern Computer

摘　　要：为了提高人脸表情识别的识别率,提出一种LBP和SVM决策树相结合的人脸表情识别算法。首先利用LBP算法将人脸表情图像转换为LBP特征谱,然后将LBP特征谱转换成LBP直方图特征序列,最后通过SVM决策树算法完成人脸表情的分类和识别,并且在JAFFE人脸表情库的识别中证明该算法的有效性。In order to improve the recognition rate of facial expression, proposes a facial expression recognition algorithm based on a LBP and SVM decision tree. First facial expression image is converted to LBP characteristic spectrum using LBP algorithm, and then the LBP character-istic spectrum into LBP histogram feature sequence, finally completes the classification and recognition of facial expression by SVM deci-sion tree algorithm, and proves the effectiveness of the proposed method in the recognition of facial expression database in JAFFE.

关 键 词：表情识别 SVM决策树 直方图序列 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]