AN IMAGE CRAWLER FOR CONTENT BASED IMAGE RETRIEVAL SYSTEM

Abstract

Images from the minute it was invented, has had an immense impact on the world we live in. The extract ing the required images from the World Wide Web (WWW) is very difficult because web contains a huge number of images. To solve this problem we need a system that can retrieve the required images needed by the user. Image Crawler is a web based tool that colle cts and indexes group of web images available on the internet. This tool collect s the keyword or phrase from the user to retrieve tfrom the web. Then these collected keyword is applied to the different general search tools like Google, Yahoo, Bind etc, . The collected web page information is stored in the temporary file till 200KB file size from the server. Then this file cont ent will be scanned and extract the image URL’s and it is compared the URL which is pre sent in the database to avoid the duplicate downloa ds. The extracted URL’s images are downloaded and finally stores unique im age and corresponding metadata like filename, url, size etc. in the database. In this paper we present the designing of an Image cra wler tool. We build a search tool which is flexible , general-purpose image search framework and explore a directed result aggregating and removing of duplicates to achieve top results compared to other existing search tools. Finally this resulted images are used in the Content Based Image Retrieval (CBIR) system for extracting the relevant images need by the client using the content of the images rather than the text based information.