Speeding-up Fractal Color Image Compression Using Moments Features Based on Symmetry Predictor

摘要

In this research, fractal compression technique using moment features based on the zero-mean range where improved by adding symmetry predictor to reduce the number of isometric trails from 8 to one trail. The first order centralized moments are used to index each domain and range blocks onto one of eight possible isometric states. At each range domain the indices of both domain and range blocks are passed through the predictor, then the predictor outputs the index of the required isometric transform (that needed to be applied on the domain block) to get the best possible match between the domain and range blocks. The moment features have been used to speed up the Iterated Function System (IFS) matching stage. These features are used to determine the block descriptor "moment''s ratio index", which in turn is utilized to classify the image blocks in both domain and range pools. During the encoding stage the block moment ratio descriptor of each range blocks is used to filter the domain blocks and keep only those blocks whose moment descriptor is suitable to be IFS matched with the tested range block. The test results showed a lower encoding time (0.16) sec with appropriate PSNR. The time of speeding is about (99.9%) in comparison with that for traditional method.