Attribute based image retrieval and hypergraph learning based image search reranking

Image search re-ranking is a powerful method to enhanced result we get from text-based search. We get noisy data from text-based search. The objective of this work is to enhance the system which re-arrange the images which user get from simple text-based search in such a way that, resultant image set contains relevant images. On this challenge, this paper proposed to use the semantic attributes for re-ranking. Every image describe through specific attribute. These attributes have already ready classifiers. User will gets responses from these classifiers on the basis of attribute. Every image in responses have relation with each other by mean of its attribute. This relation supposed to be shown by hypergraph. Images in the hypergraph ranked as per their ranking score. Ranking score represents similarity factor with respect to common attribute of images in hypergraph. This paper use attribute learning of images and hypergraph formation method to get valuable result.