Content based image retrieval based on Database revision

Abstract

Images are the simplest and best way of representation of ideas. The significance of images has been considerably increased by the web pages. Thus efficient image retrieval systems are essential. Content-based image retrieval (CBIR) systems are the latest area of research. In this paper, an intelligent image retrieval system based on a novel method called database revision (DR) is proposed. Image feature extraction in terms of color, texture and shape is employed to retrieve images from the database. The result of feature similarity comparison of the query image with database images rewrites the database. The system is made interactive for the users to identify the images that are most satisfied to the need. The user-satisfied images are analyzed and the database is revised to make the system intelligent. Experiment results and comparisons are presented to demonstrate the feasibility of the proposed method.