Image Retrieval using Bipartite Reiterative Algorithm

Content Based Image Retrieval (CBIR) depends on the visual content of an image for obtaining similar images. Research on CBIR faces a challenge of semantic gap between low level features and high level visual content. The optimal solution must reduce the computational complexity and computational time in retrieval process. Relevance Feedback is a solution to reduce the semantic gap. A new framework for content based image retrieval is proposed which is Bipartite Reiterative algorithm. This gives an approach of iteration to relevance feedback mechanism. It consists of two levels which are basic retrieval scheme and reiterative feedback scheme. Positive images are obtained from the user as part of the basic retrieval scheme. New feedback algorithm is developed based on weights assigned to visual attributes. The weights are inversely proportional to the visual attribute distance between query image and the positive images given by the user. Iterative feedback query can be performed after the user feedback. The query will be progressed according to the user feedback. The algorithm reduces the computational complexity and computational time in image retrieval.