Linear control and prediction of fractal coding time
Shengchao Zheng · Jisuanji gongcheng · 2005
Taking the Brownian dimension as their texture feature, image blocks were clustered and sorted, to control the number of domain blocks to be compared with each range block in coding. Furthermore, by combining techniques such as excluding plain blocks and building domain pool from averaged image, a fast algorithm for fractal coding was obtained, in which the coding time can be linearly controlled and predicted by the number of domain blocks compared with each range block. Experiment results indicate that with the same compression ratio, the proposed method improves the speed of coding with even better quality of decoded image, in contrast to existing classification and clustering methods.