文章基本信息

标题：Comparison And Classification Of Pictures By SOM And Metaclassifiers
本地全文：下载
作者：M.Priyadharshini ; Dr.A.Kumaravel ; Dean 等
期刊名称：International Journal of Advanced Research In Computer Science and Software Engineering
印刷版ISSN：2277-6451
电子版ISSN：2277-128X
出版年度：2013
卷号：3
期号：4
出版社：S.S. Mishra
摘要：This paper presents a technique in classifying the pictures into variety of categories or clusters desired by suggests that of Self Organizing Map (SOM) Artificial N eural Network methodology. variety of 250 color pictures to be classified as antecedently done SOMe process, such as RGB to grayscale color conversion, color bar graph, and so classifying by the SOM Feature vector choice in this paper can use 2 strategies, specifically by PCA (Principal part Analysis) and LSA (Latent linguistics Analysis) during which every of those strategies would have taken the characteristic vector of 50, 100, and150 from 256 initial feature vector into the method of color bar graph. Then the choice are processed into the SOM network to be classified into 5 categories victimiza tion a learning rate of 0.5 and calculated accuracy that is classified by the meta classifier. Classification of a number of the check results showed that the best share of accuracy obtained once victimization PCA and the choice of a hundred feature vector that's capable half of 1 mile, compared to once victimization LSA choice that 74% therefore it will be terminated that the strategy fits the PCA feature choice strategies area unit applied in conjunction with SOM And has an accuracy rate higher than the LSA feature choice strategies.
关键词：Color bar graph; Feature choice; LSA; PCA; SOM; Meta classifier