Content Based Image Retrieval System For Different Type Of Data Based On Multi-Feature Fusion Method

Abstract

The demand for the Content-based Image retrieval system is increasing these days due to the increasing interest in digital content. The process of the CBIR depends on the feature extraction process and its basis, retrieval will be done. Several Researchers proposed different feature extraction methods to improve the accuracy of the System. As features play a very important role in improving the performance so, different features can be used together to achieve the objective. To keep this in mind, a multi-feature fusion method is proposed in this paper where three different features are fused and form a single feature vector that enhances the process of retrieval. For the purpose, SIFT, SURF, and HoG features are utilized. These features are well-known features that provide information related to the shape of the object and then for matching, two different distance-based matching methods are used namely, Euclidean and Hausdrauff Distance. The experimentation results show that the performance of the proposed multi-feature fusion-based method is better than all other feature extraction methods in terms of accuracy and secondly, it is also clear from the results that Hausdorff distance method outperforms over Euclidean distance.