Speeding-Up Fractal Image Compression by Using Classification Range Blocks

In fractal compression technique, an image is partitioned into sub blocks called range blocks, each of which is encoded by matching it (after an appropriate affine transformation) with a block chosen from a large pool of domain blocks, which is constructed from the image itself. The problem is that the encoding is very time consuming because of the need to search in a very large domain pool. Our proposed approach presents a speed algorithm to reduce the encoding time called Classification Range Blocks. This technique will be reducing the size of the domain pool. The proposed method yields superior performance over conventional fractal encoding. In our proposed speeding technique, we partitioned the image by using fixed block size partitioning and computing the mean and variance for each blocks. The blocks have the variance ranging from (250, 500, 750, 1000, and 1250) only used in matching process between pair range-domain blocks.