A novel method to improve the quality of decoded images in fractal image coding

In order to improve the quality of decoded images in fractal image coding, a novel method by making use of a modified image is proposed in this paper. Since the degradation of decoded images in fractal coding mainly comes from the poor self-similarity of the input image, a modified image that satisfies good self-similarity and approximates the original image well is constructed. After transforming the input image into the modified image and performing fractal coding method on it, the decoded image with higher quality can be obtained. Simulations show that compared with the conventional fractal coding method, while the same compression ratio is maintained, the proposed method can provide higher PSNR at the expense of slightly longer encoding time.

I. Introduction

Fractal image coding is a relatively new image compression technique which has attracted much attention from the researchers worldwide[1]. As an image compression method, fractal image coding has the characteristics of long encoding time, fast decoding process and potential of higher compression ratio. In order to make it more practical, many researchers proposed their fast fractal encoding methods and the encoding process can be shortened effectively until now [2]–[5]. Moreover, other researchers focus their interest on achieving faster decoding process as well as improving the quality of decoded images [6]–[8].