Fast Fractal Image Coding Method Based on RMSE and DCT Classification
xian qiang lv,s yang,xin zhang,ying wang,y f shi,liu wei
DOI: https://doi.org/10.4028/www.scientific.net/AMM.241-244.3034
2013-01-01
Applied Mechanics and Materials
Abstract:To solve the problem of long time consuming in the fractal encoding process, a fast fractal encoding algorithm based on RMSE (Root mean square error) and DCT (Discrete Cosine Transform) classification is proposed. During the encoding process, firstly, the image is divided into range blocks and domain blocks by quadtree partition according to RMSE, then, according to DCT coefficients of image block, three classes of image blocks are defined, which are smooth class, horizontal/vertical edge class, diagonal/sub-diagonal class. At last, every range block is limited to search the best matched block in the corresponding domain block class, and the fractal coding are recorded until the process is completed. When searching the best matched block, the nearest neighbor block will be found in the sense of RMSE in the ordered codebook, and the best matched block will be further found in the vicinity of the nearest neighbor block. The experimental results show that the proposed algorithm can efficiently reduce the search space and shorten the encoding time, while achieving the same reconstructed image quality as that of the full search method.