The rapid expansion in web environment and advancement in technology have led us to access and manage enormous images easily in various fields. Present internet image search engines purely faith on keyword based information over images. Keywords provided by user cannot determine content of images correctly. Resultant images comprised of several disordered and uncertain images. To clarify this doubtfulness in keyword-based image search, it is valuable to use visual details of image. System has been designed to overthrown unpredictability of images such that users aim can be determined by one click internet Image search. User preferred a query image in set of images returned by expanded keyword-based search. This technique provides extra clusters generated by applying combination of candidate words and visual features of images by Scale Invariant Feature Transform (SIFT) algorithm. Weight of image is calculated by providing procedure of weight adaption. Results are refined by re-arranging of images by homogeneity calculation. Message Digest (MD5) hash function is used for detecting and removing duplicate images. Resolution of images is considered for further enhancing quality of images. Intention of user is estimated by combining textual and visual resemblance without utilizing extra efforts. An experimental observation determines expressive enrichment in concern with user justice and significancy.
Keywords
text based search, visual features, adaption weight, cluster.
