期刊名称：International Journal of Computer Trends and Technology
电子版ISSN：2231-2803
出版年度：2018
卷号：55
期号：1
页码：31-35
DOI：10.14445/22312803/IJCTT-V55P106
出版社：Seventh Sense Research Group
摘要：In past recent years, Locality preserving Projection (LPP) has proved to be an alternative to Principal Component Analysis in Face Recognition. Despite the fact that LPP is better than PCA, it has some limits. In order to overcome that limits, many new methods still emerging. In this paper we propose a method using Locality preserving Projection on Wavelet Subband for features extraction and Artificial Neural Network for recognition. A comparative study has been done between the new method, the Locality Preserving Projection, the Principal Component Analysis and the Principal Component Analysis on Wavelet Subband by considering execution time, recognition rate and dimension reduction power. Experiments have been done on two face data bases ORL and Yale data bases. Results show that the new method improve a little bit the execution time.
