Effectiveness of relevance feedback for content based image retrieval using Gustafson-Kessel algorithm

Abstract

The performance of the Content Based Image Retrieval (CBIR) can compute using similarity of the images where user can retrieve from the image database. The term similarity in the mind of the user may different depends on the search query and the experience of the user which has been using the similar applications. When the users are not satisfied with their search results, the relevance feedback (RF) retrieval is one of the solutions for this critical problem. The user needs to use this feedback on the next retrieval process in order to increase the retrieval performance. In this paper, we have used a relevant feedback approach based on Gustafson-Kessel (GK) clustering approach in order to evaluate the effectiveness of the image retrieval results from the users. From the experiments, we have found that the RF method using Gustafson-Kessel (GK) clustering can improve the retrieval performance of the CBIR system even if we are using a large set of image datasets with a variety of images.