A Novel Compression Technique for Color Image Database

Abstract

This paper presents a novel compression technique for color image database using fractal coding and spiral architecture. Fractal coding is used for gray level images. Although new approaches have been proposed that compress gray level and color images, the proposed approach, firstly, forms one-plane image from three color planes of RGB color space image. Then the one-plane image in traditional square structure is represented in spiral architecture for compression. On this new spiral architecture image, fractal coding is applied to get encoded image. Extensive experiments are carried out with various tolerance values. Median is used as the basis in codebook formation. Implementation results are compared and show more compression with little trade off in image quality. Due to the spiral multiplication and minimization of color planes from three to one, the proposed approach is more effective. Proposed approach falls in lossy compression category, with time complexity in the order of n i.e. O(n).