结合差图像和Gabor小波的人脸表情识别  

FACIAL EXPRESSION RECOGNITION BASED ON COMBINATION OF DIFFERENCE IMAGE AND GABOR WAVELET

作　　者：丁志起[1] 赵晖[1] 

机构地区：[1]新疆大学信息科学与工程学院,新疆乌鲁木齐830046

出　　处：《计算机应用与软件》2011年第4期47-49,共3页Computer Applications and Software

基　　金：国家自然科学基金(60962005);新疆维吾尔自治区自然科学基金(200821124);教育部科学技术研究重点项目(208164)

摘　　要：提出一种结合差图像和Gabor小波变换的人脸特征提取方法,并使用支持向量机SVM(Support Vector Machines)进行人脸表情识别。对包含情感信息的静态灰度图像进行预处理,将眼睛和嘴巴等表情子区域从人脸中切割出来,求出其差图像,然后提取差图像的Gabor特征,使用下采样降维减少特征向量的维数并进行归一化,最后使用SVM进行分类。与只从表情子区域提取Ga-bor特征的识别方法进行了比较,结果显示识别效果更好。In this paper we introduce a facial expression features extraction algorithm which is the combination of difference image and Gabor wavelet transform,and use the support vector machine（SVM） to recognise facial expression.For a given static grey image containing facial expression information,pre-processing is executed first,the expression sub-regions including the eyes and the mouth respectively are cut from the face for obtaining their difference images,then we extract Gabor feature vectors of the difference images,and employ downsampling to reduce the dimensionality of the eigenvectors,and normalise the treated data,finally we use SVM to classify the facial expression.This combination method has been compared with the recognition method which only extracts the Gabor feature from expression sub-region,the result indicates that the combination one has better recognition performance.

关 键 词：人脸表情识别(FER) 差图像 GABOR小波 支持向量机(SVM) 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]