非负矩阵分解方法是基于局部特征的特征提取方法,已经成功用于人脸识别。研究基于非负矩阵分解的人脸图像识别的改进算法是一个有重要意义的研究课题。采用二维非负矩阵分解方法(2DNMF)和对角非负矩阵分解方法(DiaNMF),并且使用正交的基矩阵进行Matlab实验。实验结果表明,以上改进措施能够有效提高人脸图像识别的正确率。
Non- negative matrix factorization(NMF)is a method of parts- based feature extraction, it has been already applied to face recognition successfully. It is an important issue to research the improved method in face recognition field. In this paper,2 - D non - nega- tive matrix factorization and diagonal non- negative matrix factorization are adopted, and use orthogonal base matrix to make experiment. The experimental result shows that, compared with developed face recognition based on non- negative matrix factorization, the improved algorithm can increase accurate ratio of face recognition.