一种基于2D-DWT和2D-PCA的人脸识别方法

Face recognition based on 2D-DWT and 2D-PCA

提出了一种联合图像二维离散小波变换(2D-DWT)和二维主成分分析(2D-PCA)的人脸识别方法。首先通过2D-DWT将当前图像分解成四个子图像,其中一子图像对应原图像的主体部分(低通部分),其余三个子图像则对应图像的细节部分(高通部分)。在此基础上,采用2D-PCA方法分别对每一子图像进行特征提取。此外,文中还提出了一种简单有效的方法对各子图像中所提取的特征进行融合,根据所得到的特征进行人脸识别。同其他基于小波分解的人脸识别方法相比,所提出的方法能更充分地利用人脸图像的有用判别信息,并得到更好的识别结果。

An efficient face recognition method by combining the 2D-DWT（two-dimensional discrete wavelet transform） method with the 2D-PCA（two-dimensional principal component analysis） method was proposed. First, each face image was decomposed into four sub-images by using the 2D-DWT approach, and then 2D-PCA approach was used to extract the features for recognition from each sub-image respectively. All the extracted features were further combined and used for face classification. Moreover, considering that the discriminative features extracted from each sub-image may not share the same metric scale measure, we also proposed an effective features combination method in this paper. Better performance of the proposed method is confirmed by the Yale face database and the AR face database.