Fractal Image Compression with Fast Local Search

摘要: Optimal fractal image compression is an NP-hard combinatorial optimization problem where the domain of feasible solutions a large finite set T contractive affine mappings, and cost function \( ||{f^{*}} - {f_{T}}||_{2}^{2} \) f* original image, f fixed point G T. In contrast, traditional coders are based on greedy algorithm known as collage coding, which minimizes T\left( {{f^{*}}} \right)||_{2}^{2} \). We describe local search that rapidly improves solution obtained by coding. particular, we show how successive computations can be efficiently done combining Gauss-Scidel like iterative method graph algorithm.