基于奇异值与稀疏表示的稳健性人脸识别

Robust Face Recognition Based on Singular Value and Sparse Representation

目前基于二维的正面人脸识别面临着两大难题,一是光照、表情变化问题,二是遮挡、噪声问题。这些正是研究人脸表情、光照、遮挡和噪声变化的稳健性人脸自动识别问题。将识别问题看作是多个线性回归模型中的分类问题,并用稀疏表示理论解决这些问题。基于奇异值与稀疏表示理论可以显著提高对噪声和遮挡变化的稳健性并降低计算复杂度。在公用的人脸数据库上进行实验,并证实算法的有效性。

At present, there are two main barriers toward positive two-dimensional-based face recognition： one is the variation of illumination and expression, the other is the problem of occlusion and noise. The problem of robust recognizing human faces from frontal views with varying expression and illumination, as well as occlusion and noise will be researched in this paper. The recognition problem is taken as one of classifying among multiple linear regression models, and sparse signal representation is used to solve this problem. The method based on singular value and sparse representation can significantly improve the robustness and reduce the computational complexity. Conducting extensive experiments on publicly available databases verify the efficacy of the proposed algorithm, and corroborate the above claims.