Using adaptive contraction for fractal image coding based on local fractal dimension

In automatic fractal image compression, most of the encoding time is spent on finding the best match between each range block and the domain blocks. We propose the use of the local complexity of the image domain blocks to reduce the number of pairs to be tested on this search. Indexing the contracted domain pools according to their local fractal dimension and using multiple contraction factors we can confine all potential matches to a relatively small number of possibilities. By selecting an appropriate criterion for close match, the compression time can be shortened without image quality degradation. The performance of the proposed algorithm, evaluated by means of fidelity versus encoding time and compression ratio, is compared with five approaches. Future developments can also improve the compression ratio by using fractal dimension in the definition of the size of range blocks.