基于Uniform LBP和DMMA的单样本人脸识别

在单样本人脸识别系统中，为了获得更好的人脸面部特征，提出了一种融合Uniform LBP特征和多流形判别分析（Discriminative Multi-Manifold Analysis，DMMA）的特征提取方法。对每幅人脸图像进行分块构成一个子集。使用统一局部二值模式（Uniform LBP）算子提取每个子集中图像的直方图，每个子集中的直方图形成一个统计流形，应用DMMA算法获得人脸图像的低维特征。采用基于重建的流形-流形间的距离识别未知的人脸图像。在AR数据库和ORL数据库上实验结果表明，该算法的识别性能优于一般的DMMA算法。

Face recognition with single training sample per person based on Uniform LBP and DMMA

In order to effectively extract facial expression feature in the single face recognition system, this paper proposes a novel method by fusing Uniform LBP features and  Discriminative Multi-Manifold Analysis（DMMA） features. Each face image is partitioned into several nonoverlapping patches to form an image set for each sample per person. The Uniform Local Binary Pattern（Uniform LBP） operator is used to extract image histogram of each image set. So the histogram of each image set forms a statistics manifold. This paper applies DMMA algorithm to obtain the low-dimensional face image feature. It uses the reconstruction-based manifold-manifold distance to identify the unlabeled face images. Experimental results show that the algorithm is superior to the general recognition DMMA algorithms on the AR database and ORL database.