Image compression based on fractal theory is known to have high compression rate and fast decoding time. However, it is limitedly used in practice since it spends too long time to find matching blocks in encoding process. In other to reduce the encodi...
Image compression based on fractal theory is known to have high compression rate and fast decoding time. However, it is limitedly used in practice since it spends too long time to find matching blocks in encoding process. In other to reduce the encoding time we propose a new method to reduce the number of comparisons using the isometry transformation. The proposed method classify the blocks into one of four types such as plain, middle, vertical/horizontal, or diagonal block. The block search time can be reduced by applying the 8 isometry transformation process in different ways depending on the block types. The experiments show that the proposed method reduces the compression times up to minimum 35% smaller than those of 8 isometry transformations while maintaining PSNR of 27∼30dB.