摘要
研究人脸表情识别问题,应有效提取脸表情特征,消除与识别无关的信息。传统的Gabor滤波器在人脸表情特征提取过程中,针对存在提取特征时间较长和特征数据存在冗余性的缺点,提出了一种Gabor和PCA相结合的特征提取,并通过支持向量机进行表情识别方法。方法首先对人脸表情进行预处理得到纯表情图像,采用Gabor提取表情特征,用PCA进行数据冗余处理和用支持向量机识别人脸表情并进行仿真。仿真结果表明,相对于传统的Gabor方法,不仅提高了人脸表情识别的正确率,而且加快了识别的速度。改进方法非常适合于人脸表情图像的分析。
Facial expression recognition problem is studied.The traditional Gabor filter needs a lot of time to extract feature vectors and the feature vectors are very redundant.A facial expression recognition method is proposed based on Gabor and PCA,and then is recognized by support vector machine(SVM).Face facial expression image is pretreated firstly,the feature is extracted by Gabor and the redundant message is filtered by PCA,finally facial expression recognition is carried out by SVM.Experimental results show that compared with traditional methods,this method not only improved the facial expression recognition accuracy,and accelerates the speed of recognition.This method is suitable for facial expression recognition of image analysis.
出处
《计算机仿真》
CSCD
北大核心
2011年第2期337-340,共4页
Computer Simulation
基金
2009年广东自然科学基金项目(9151027501000039)
湛江市科技攻关项目(2007C09017)
关键词
人脸表情识别
特征提取
主成分分析
支持向量机
Facial expression recognition
Feature extraction
PCA
Support vector machine(SVM)