Image Searching Using Client Server Architecture

Scalable image search based on similarity matching has been an active topic in recent years. Currently use of web has been increased significantly for information recovery and it is challenging to extract the relevance information in less time. For this the State-of-the-art methods usually use hashing approaches to embed high-dimensional image features into given Hamming space, where result search may be executed in real time based on Hamming distance of compact binary hash codes. There are various methods based on account of query adaptive method to recover the image searching. Now days, with the increasing scale of image databases, incorporated image recovery system no longer provide sufficient quick search. In this, we work on a scalable client server architecture, which is equivalent to web search engine, for effective large scale image retrieval. In our client server architecture, images are partitioned to multiple clients and an index is built. Administrated by a controlling one server, each client matches query image in its own image sub-collection in parallel and returns the intermediate search results, a list of images related to query image. An evaluation of the results shows that our client server architecture removes the limitation of an integrated image recovery system and this can produce better results. Analysis on a Flickr image dataset and relevance feedback with client server architecture for given output illustrates perfect improvements from our projected approach.