Fast Fractal Image Encoding Based on Local Variances and Genetic Algorithm

Fractal image encoding is computationally expensive and consumes longer time, which limits its workable applications. This paper proposes an improved method. Firstly the variances of the normalized image blocks are regarded as the classification features and are used to classify both the original range and domain blocks into six classes. Secondly this paper utilizes the genetic algorithm as the search strategy to search the similarity matching blocks in the same classes of the corresponding ranges. The experiments results show the validity of the presented approach in accelerating fractal encoding process and in holding the quality of the decoding image.