Heterogeneous Sim-Rank System For Image Intensional Search

Abstract

An image-rich information network is a social media site for uploading images by users which are associated with information about user, consumer producer, annotations .It uses a combined approach which measures the similarity based on both link based and Content based. The link based similarity depends upon the social network information like tags, groups and annotation over the images .Content based similarity considers the image content properties edge, color histogram, texture shape etc. Then, by considering the network structure and reinforcing link similarity use an algorithm Integrated Weighted Similarity Learning (IWSL) to find both link-based and content based similarities. The combination of two methods to integrate the social resources and helps to classify the images in image-rich information networks. It implements a new search engine system to find relevant products from online media