Non-Exhaustive Search Method for Fractal Image Compression

Encoding images based on self-similarity property is one of the essential coding methods used for texture and natural image compression. The two important considerations of fractal compression are higher compression ratio and quality preservation, but this technique faces a large encoding time, each range block must have a well-matched contracted domain block from a huge pool of blocks, which needs a large search time. To optimize the searching time required for the encoding process with improvements of both fidelity and compression ratio, we construct a flexible domain pool encoder based on image segmentation techniques combined with fractal coding. A comparison result between recent FIC algorithms and the proposed one is introduced.