融合半监督降维与稀疏表示的人脸识别方法

由于人脸图像数据的维数都较高,将稀疏表示分类用于人脸识别时计算量很大,为了提高人脸识别系统的效率,提出了一种融合半监督降维和稀疏表示的人脸识别方法。首先利用半监督降维算法对图像进行降维处理,在较低的维数空间快速取得较高的识别率,然后利用稀疏表示分类进行人脸识别,取得比传统的最近邻分类器更高的识别率,最后在ORL人脸库上进行实验验证。结果表明,利用该融合算法可快速有效地提高人脸图像的识别效果。

Face recognition based on semi-supervised dimensionality reduction and sparse representation

Because of high dimensions of face image data and large calculation of sparse representation classification for face recognition, in order to improve the efficiency of face recognition system, a new face recognition method based on semi-supervised dimensionality reduction(SSDR) and sparse representation (SR) was proposed. Firstly, SSDR algorithm was used to reduce the image dimensions and achieve higher recognition rate in the lower dimension space quickly. Secondly, SR classification can achieve a higher recognition rate than the nearest neighbor classification in face recognition. And then, the experimental verification was demonstrated on ORL face database. The results show that the fusion algorithm can improve the recognition performance of face images quickly and effectively.