文章基本信息

标题：Performance study of Face Recognition systems using LBP and ICA descriptors with sparse representation - MRLSR and KNN Classifiers, respectively
本地全文：下载
作者：K Sarath ; G. Sreenivasulu
期刊名称：International Journal of Computer Trends and Technology
电子版ISSN：2231-2803
出版年度：2016
卷号：42
期号：1
页码：33-41
DOI：10.14445/22312803/IJCTT-V42P106
出版社：Seventh Sense Research Group
摘要：In this paper an attempt has been to study and compare the results of two scenarios, i.e. 1) LBP descriptor with sparse representation (MRSLR) classifier and 2) ICA descriptor with KNN classifier, of a face recognition system, on a standard image data sets for training and testing with/without noise, occultation and different illumination facial images. The results show that the second scenario ICA + KNN exhibits better performance, compared to the first scenario LBP + MRSLR as descriptors and classifiers, respectively. The results show that the ICA + KNN scenario exhibits better performance, with a faster recognition speed and more recognition accuracy.
关键词：Local Binary Pattern (LBP); IndependentComponent Analysis (ICA); Sparse Representation; MRLSR; KNN