一种基于离散小波变换和支持向量机的人脸识别新方法

为提高人脸识别系统的性能，提出了一种基于离散小波变换DWT（discrete wavelet transform）特征提取和支持向量机（SVM）分类的人脸识别方法。首先，采用DWT对人脸图像进行降维和去噪，然后，对小波低频子图像进行核辨别分析（KDA）提取人脸特征，最后，结合SVM进行分类识别。基于该方法，对ORL人脸库进行分类识别，采用39个特征识别率达到98．2％。仿真结果表明，该方法明显减少了高频干扰对人脸特征的影响，增强了特征的辨别能力。而且，SVM有效地提高了分类器的分类和推广能力。

Novel face recognition method based on DWT and SVM

To improve the performance of face recognition system,a novel face recognition method based on discrete wavelet transform(DWT) and support vector machine was presented.The raw face images were denoised by the DWT at first.Then the kernel discriminant analysis(KDA) was performed on the waveletfaces to enhance discriminant power.Finally,the support vector machine(SVM) was selected to perform face classification.Experimental results on ORL face database show that the proposed method achieves a recognition accuracy of 98.2% using only 39 features.