期刊名称：International Journal of Innovative Research in Science, Engineering and Technology
印刷版ISSN：2347-6710
电子版ISSN：2319-8753
出版年度：2014
期号：ICETS
页码：384
出版社：S&S Publications
摘要：Mining of Structured representations incontent based image retrieval is a popular researchtopic in many useful applications. In the last decadethere has been an explosion of interest in miningtime series data. Literally hundreds of papers haveintroduced new algorithms to index, classify, clusterand segment time series. The initial work focusedmainly on values with tags, while most of the recentdevelopment focuses on discovering association ruleamong tree structured data objects to preserve thestructural information. In this paper we combinedthe techniques of texture based segmentationalgorithm, Blob reduction by identifying outlier’sdetection, SIFT algorithm to an automatic system toannotate and retrieve images. This paper tend toreveal a good behaviour in classification of ourgraph based solution on two publicly availabledatabases and produce the images features withmore enhancement by an efficient segmentationalgorithm.
关键词：Scale invariant feature transform;(SIFT); Blob Reduction and Extraction;Segmentation; Global method; key point descriptor
