Fractal Image Coding Based on High Order Spectrum Using Nonparametric Estimation

Abstract

This paper presents a new strategy to speed up the encoding process for fractal image compression. First, the range blocks are obtained by partitioning the image using adaptive quadtrees. Then we extract the high-order spectrum based on nonparametric double-spectrum estimation to constitute the eigenvector for the image block which size is larger than 4times4. The lower dimension kd-tree structure is to be created while orthogonal projecting these eigenvectors of high-order spectrum for the image block. The fractal code is quickly obtained by using nearest neighbor searching algorithm and quantizing the transform parameters. The experimental results show that, in comparison with conventional methods, the proposed algorithm can provide better speed-up and image quality under the same compression ratio.