Combining visual features of an image at different precision value of unsupervised content based image retrieval

Abstract

Image Retrieval is a technique of searching, browsing, and retrieving the images from an image database. There are two types of different image retrieval techniques namely text based image retrieval and content based image retrieval techniques. Text-Based image retrieval uses traditional database techniques to manage images. Content-based image retrieval (CBIR) uses the visual features of an image such as color, shape, texture, and spatial layout to represent and index the image. CLUE (CLUster based image rEtrieval) is a well known CBIR technique retrieves the images by clustering approach [4]. In this paper, we propose a CBIR system that also retrieves images by clustering just like CLUE. But, the proposed system combines all the features (shape, color, and texture) with some percentage of all features value for the purpose. In this paper we proposed two methods of CBIR by combining some percentage value of two features namely color-texture features and color-shape features and we also take the union of these two features. This combination of features provides a robust feature set for image retrieval. We evaluated the performance of proposed methods at different precision value of the image retrieval on each category of image database. We compared the performance of the proposed system with the two other existing CBIR systems namely UFM and CLUE at precision 100. We experimented with COREL standard database of images and experimentally, we find that the proposed system is better than the other two existing systems, and at smaller precision value result outperforms in almost each category of images.