支持向量机在人脸识别中的应用研究

Study about Face Recognition Based on Support Vector Machine

研究人脸图像自动识别问题,由于人脸的特征维数较高,正确识别有难度,利用计算机技术对人脸图像进行分析,从中提取有效的特征来识别出人的身份,其关键技术在于人脸特征的提取和模式识别。为此,提出了一种基于支持向量机的人脸识别方法。首先采用Gabor滤波器提取人脸图像的特征,PCA降维处理消除人脸特征之间的冗余信息,然后采用支持向量机对提取特征进行训练得到最优识别模型,用最优模型对人脸进行识别。对ORL人脸图像库进行仿真实验,识别率达98%,比传统算法高出5%,实验结果表明,既减少了计算复杂度,降低训练与识别时间,又保证实时性,提高识别正确率,为人脸识别的应用提供广泛的前景。

Study about face recognition problem.The accuracy and efficiency of face recognition algorithm for traditional identification are low,is low.A new method of face recognition was proposed based on support vector machine.Firstly.The features of face characters were detected by Gabor filter and the principal component analysis feature extraction.Then the features were used to train support vector machine classifier.Finally,the face characters were classified by the support vector machine.The ORL face database was used to test the proposed method.Experimental results show that the recognition accuracy is above 98.0% and is by 5% higher than traditional algorithm.This method reduces not only the computational complexity and the training and recognition time,but also ensures the accuracy of real-time,improve recognition.This method in face recognition has been widely applied prospects.