An efficient fractal-based algorithm for image magnification

Abstract

An efficient image magnification algorithm based on the Iterated Function System (IFS) is proposed in this paper. This IFS-based image magnification method employs the self-similarity property instead of the conventional interpolation approach. This self-similarity makes it possible to generate images of higher resolution. Our proposed algorithm can maintain not only the high frequency information that preserves the edges, but also effectively remove blocky artifacts from a magnified image. To further improve the quality of the high-resolution images, the error image or residual errors are considered. In addition, our algorithm can combine with other magnification algorithms. Experimental results show that our proposed algorithm can increase the PSNR by 2.60dB, 2.30dB and 0.83dB on average when compared to the traditional fractal-based image magnification algorithm, the bicubic spline interpolation method and the recently proposed fractal-based image magnification algorithm (IEUF), respectively. If residual errors are considered, the PSNR of our algorithm can be further increased by about 0.22dB. When combining our algorithm with IEUF, the PSNR can further improve by 0.4dB, on average.