CONTENT BASED IMAGE RETRIEVAL SYSTEM USING SHAPE TEXTURE AND COLOR FEATURES

Abstract

Content based image retrieval is an important field of multimedia applications. This has been used for extraction of similar images from the large datasets. Systems that have been used for content based image retrieval system use various features that are low level. Various approaches and features have been widely used for extraction of relevant information. In this paper a new approach has been proposed that use combination of colours, shape and texture features. SURF and Fuzzy has been combined in this system for efficient retrieval of the relevant images. SURF is a fast approach that compute point based descriptions about the shape features of the images. Fuzzy interference has been used for computation of fuzzy membership values so that efficient feature vector can be developed. On the basis of parameters analysis we can predict that proposed approach provide better performance.