Colour Based Information Retrieval

Abstract

The problem of retrieval and organization of the visual information of images is a crucial one for specialists, since the visual information processing requires a perceptual ability that is not yet formulated into computational model. Therefore the ability to retrieve the visual information is a complex problem. The retrieval pattern mainly includes the features of an image such as color, shape, number of regions, edges and so on. Hence, algorithms are developed to compute similarity between the input image and an image in the database based on the color features to compare and retrieve the image content. In this paper an attempt is made to compute similarity of images based on color regions and spatial information of red, green and blue (RGB) present in an input image for retrieving the similar images from database. Experiments are carried out with proposed algorithm and obtained an accuracy of 93.53%.