Distance measures for colour similarity, Human perception methods for Image Matching, analysis and Retrieval

Abstract

Image processing is needed for major applications like improving pictorial information for better human perception and autonomous machine applications. In this regard the information is first converted into discrete and several actions are performed and different algorithms are applied to perform digital image processing. Color speaks more information regarding a picture than an black and white picture. the proposed methodology defined here is the color theory. Colour descriptors play a major role in image processing, analysis and retrieval systems. Colour descriptors are preferred due to their low complexity and compact representations. Here we compare colours with histogram comparison. Even though it is having so many disadvantages like sensitivity to quantitization boundaries and need of colour codebook design with high dependence. It is also having drawback of representing image with few dominant colour. Here we represent an efficient algorithm for colour matching that models the features and working of human visual perception for capturing colour representation in an image. Here we present optimal methods for mapping sets of colour components representing images. Our approach leads to better match of similarity judgment on perceptual dimensions (human perception) has to be modeled along and simulations in the image colour composition.