摘要
人脸识别是利用计算机提取人脸的相关特征,并由此辨别人物身份的一种应用技术,人脸图像的特征提取是人脸识别过程中最关键的技术之一,局部特征提取方法是人脸特征提取方法中比较常用的一种方法,而非负矩阵分解算法是一类应用较为广泛的局部特征提取方法。以人脸识别中的特征提取作为研究对象,在已有的非负矩阵分解算法理论的基础上,将矩阵变换理论、最速下降法和稀疏性理论结合,分别提出了基于矩阵变换的非负矩阵分解算法和基于最速下降法的非负矩阵分解算法,将改进的算法应用到人脸识别中,并讨论改进的算法与传统非负矩阵分解算法在人脸识别中的不同效果。结果表明,提出的算法相对于传统的NMF算法具有运行时间快、误差小的优点。
Face recognition is used to extract the relevant features of a face by the computer,and thus to be an applied technology for identifying persons.Face image feature extraction is one of the most key technology in the process of face recognition,and the local feature extraction method is more commonly method in face feature extraction method.While Non-negative Matrix Decomposition is a widely used method for local feature extraction.The feature extraction of face recognition is used as the research object,on the basis of the existing non-negative matrix decomposition algorithm,combined with the matrix transformation theory and the steepest descent method and the sparse theory,so the non-negative matrix decomposition algorithm based on the matrix transformation and the non-negative matrix decomposition algorithm based on the steepest descent method are proposed respectively.And the improved algorithm is applied to the face recognition.The results show that the proposed algorithm has the advantages of fast running time and small error compared with the traditional NMF algorithm.
作者
彭梦冉
PENG Meng-ran(School of Information Engineering,Anhui Business and Technology College,Hefei 231131,China)
出处
《长春工程学院学报(自然科学版)》
2019年第3期104-108,共5页
Journal of Changchun Institute of Technology:Natural Sciences Edition
关键词
人脸识别
非负矩阵分解
最速下降法
稀疏约束
face recognition
non-negative matrix decomposition
steepest descent method
sparse constraint