期刊名称：International Journal of Advanced Research in Computer Engineering & Technology (IJARCET)
印刷版ISSN：2278-1323
出版年度：2012
卷号：1
期号：10
页码：14-18
出版社：Shri Pannalal Research Institute of Technolgy
摘要：This paper presents an image retrieval system in a computer system for browsing, searching and retrieving images from a large database of digital images. The support vector machine (SVM) is a concept based on learning methods that evaluate data and identify patterns, used for classification and regression analysis. In this work, the prediction of database and query image depends on the outputs of a multiclass support vectormachine (SVM) classifier. The irrelevant images are first to filter out, for similarity matching. The different low-level featuresmodalities are used in this framework. When compare to the query image, the main difficulty is to find unique feature. So an adaptive linear combination of similarity matching and user's feedback information is used here. The predicted output of classifier may be inexact in some cases. Hence, the feature weights are calculated by considering both precision and rank order information based on users feedback. As a result, using SVM classifier with the retrieval performance clearly shows the advantageof searching images based on similarity fusion and filtering in terms of competent and success.
