Integrated algorithm (S-CBIR) for image retrieval in imase-rich information networks

Social networking sites like Facebook and Flickr allows users to upload billions of images, famous e commerce web sites such as flipkart, amazon etc. are also provides number of product related images to user. Images which are present in social multimedia networking websites are accompanied by different tags, comments, annotations and other related information which tend to form an image rich information networks. Due to existence of different related information to images (tag, annotation etc.) retrieval of images from image rich information networks is very challenging and complex task. In this paper we have proposed an integrated approach (S-Cbir) to retrieve an images from image rich information networks. The proposed algorithm consists integration of two different techniques such as link based image retrieval and content based image retrieval systems. K-Simrank algorithm is first applied to find out the node similarity in image rich information networks and then CBIR technique is applied to find out relevant images. SIFT, CEDD, feature and texture extraction descriptors are used in CBIR technique. The search results shows that proposed algorithm refines the search results and also improves the accuracy of search.