Text Based Web Image Retrieval System by Using Latent Semantic Indexing (LSI)

Abstract

This paper describes the use of latent semantic indexing (LSI) in text-based image retrieval system. Due to the rapid development and popularity of World Wide Web (WWW), users have to face a variety and large number of web pages and images that waste time to search and browse. To alleviate this difficult, more efficient image retrieval system is needed to extract the required image from the collection of web pages. Therefore many image retrieval systems have been developed and used on the web. Among them, text-based image retrieval is a kind of image Meta search that based on associated metadata. Our proposed text-based image retrieval can retrieve both images alone and combined of images and associated alphanumeric text.