Image coding using a fractal/vector quantization model

Abstract

In this paper a novel image coding scheme, based on coinbination of fractal and vector quantization techniques is proposed. A two-level coding scheme is described. In the first level, a form of vector quantization is employed to get a rough first-order approximation of the image. In the second level, the residual image, produced after subtracting thc first-order approximation limn the original image, is coded using a simple, fixed-block size fractal technique. Using fixed block size, the algorithm has a very simple and regular structure. The benefit of the proposed technique is the reduction of' the computational complexity in comparison to traditional fractal coding, while the image quality is kept i n high levels

Gradient match and side match fractal vector quantizers for images

“…The Lloyd algorithm for VQ can be used to reduce the number of domain blocks such that the lower encoding complexity can be achieved [5]. The mean shape-gain vector quantization (MSGVQ) is combined with the fractal image compression in [3].…”