摘要： 针对人脸识别对遮挡、表情和光照的鲁棒性问题，提出基于PCA特征基压缩传感算法的人脸识别方法。利用双向二维主成分分析提取图像行列2个方向的特征并进行降维，建立反映人脸特征投影矩阵，作为压缩传感算法的超完备基。通过求解最小化l1范数，寻求图像在该超完备基上的稀疏表示，以得到一组最优稀疏系数重构各类图像，求取测试图像与各类重构图像的最小残差进行分类识别。实验结果表明，该方法在较低的人脸特征维数下具有较高的人脸识别率，能有效提高人脸识别对遮挡、表情和光照的鲁棒性。

关键词: 人脸识别, 压缩传感, 稀疏表示, 最小化l1范数, 鲁棒性

Abstract: This paper presents a face recognition method based on Compressed Sensing(CS) algorithm on PCA feature set for robustness problem caused by occlusion, expression, as well as illumination. It utilizes the Two Directional Two Dimensional PCA((2D)2PCA) transformation to extract image features in both row and column directions and reduce the dimension. A projection matrix is constructed to identify the face features, considering these features to form an over complete dictionary. By solving the l1 norm minimization, the paper finds out the sparsest representation of images based on the dictionary to obtain a set of optimal sparse coefficient, which is used to recover the train images, computes the residuals between test and train images for face recognition. Experimental results show that this method not only has a high recognition rate in a lower dimension, but also reduces the computational complexity. Thus it can effectively improve the face recognition robustness on occlusion and expression, as well as illumination.

Key words: face recognition, Compressed Sensing(CS), sparse representation, minimization l1 norm, robustness

中图分类号: 

• TP18