Image Retrieval Using Bat Optimization and Image Entropy

Abstract

Technological advancements have increased the size of image databases and their applications in different fields. This has also increased the importance of image processing and its subfields like image classification, image segmentation, image retrieval, image enhancement, image compression, image restoration etc. This is also the reason behind the increase in the number of research works on image retrieval. It has inspired people to develop new techniques on image retrieval. Text-based retrieval techniques were gradually replaced by content based image retrieval techniques. Content based image retrieval techniques primarily based on the retrieval of images based on primitive features like colour, shape, texture etc. The proposed method considers image entropy and the Bat algorithm for image retrieval. Entropy helps to find the degree of randomness in the images which can be analysed to obtain their similarity. Bat algorithm is applied to obtain optimal values based on which most similar images are retrieved. For experimental analysis of the proposed technique, both medical and nonmedical images are considered. The results obtained prove the effectiveness of the proposed approach in the retrieval of different categories of images.