Improving the Performance of Content Based Image Retrieval System using Relevance Feedback

The performance of the content based image retrieval (CBIR) system can be improved by reducing the semantic gap between visual features and human semantics. Relevance feedback (RF) approach refines the retrieval process with users feedback on CBIR system results. A variety of RF methods have been widely used to reduce the semantic gap. It was observed that existing RF techniques face the challenges of number of iterations and the execution time. To improve the retrieval efficiency of the existing system, the proposed RF approach uses classification based method. The positive and negative examples provided by the user will be used for the classification. A binary classifier will be trained to distinguish between relevant and irrelevant images according to the preferences of the user. The trained classifier will be later used to provide an updated ranking of the database images represented in the space of the selected features.