基于判别超图和非负矩阵分解的人脸识别方法

非负矩阵分解是一种流行的数据表示方法,已广泛应用于图像处理和模式识别等问题.但是非负矩阵分解忽略了数据的几何结构. 而现有的基于简单图的学习方法只考虑了图像的成对信息,并且对计算相似度时的参数选择非常敏感. 超图学习方法可以有效地解决这些问题. 超图利用超边将多个顶点相连接用以表示图像的高维结构信息. 然而, 现有的大部分超图学习方法都是无判别的学习方法.为了提高识别效果, 提出了基于具有判别信息的超图和非负矩阵分解方法的新模型, 利用交替方向法进行迭代求解新模型, 并结合最近邻方法进行人脸识别. 在几个常用标准人脸图像数据库上进行实验, 实验结果表明提出的方法是有效的.

A face recognition method based on discriminative hypergraph and nonnegative matrix factorization

Nonnegative matrix factorization (NMF) has become a popular data representation method and has been widely used in image processing and pattern recognition problems. However, NMF ignores the local geometric structure of data. Existing simple graph-based transductive learning only considering the image information in pairs, and they are sensitive to the radius parameter used in similarity calculation. Hypergraph learning has been investigated to solve the problems. Hypergraph models the high-order relationship of samples by using the hyperedges to link multiple samples. However, most of the existing hypergraph learning methods are unsupervised methods. Based on the discriminative hypergraph and nonnegative matrix factorization, we propose a new model and solve the new model by using the alternating direction method of multipliers. The new method, together with the nearest neighbor method, is applied to face recognition. Experimental results on several standard face datasets show the effectiveness of our method.