基于图像预处理下的2DPCA及人脸识别

对人脸图像进行离散小波变换来消除部分对识别无关的信息,以提高识别率并有效降低时间复杂度。同时为了抑制光照等外界条件的负面影响,还引入一种对图像灰度进行指数衰减的预处理策略,对预处理后的人脸图像采用二维主成分分析方法进行人脸识别。在YALE和ORL人脸库上试验表明,结合图像预处理的二维主成分分析(2DPCA)方法有着比PCA以及2DPCA更好的识别效果。

2DPCA and Face Recognition Based on Image Processing

This paper eliminates the irrelevant factors through discrete wavelet transformation on human face images to im- prove the recognition rate and reduce the time complexity. In order to reduce the external negative effects like light and illu- mination,the study also introduces a strategy to preprocess the image gray via exponential decay. The processed human faces are recognized through 2DPCA. The experiments on the YALE and ORL face database show that combination of image pre- processing and 2DPCA is much better than PCA and 2DPCA in terms of recognition effect.