Fractal Image Compression Mechanism by Applying Statistical Self-Similarities
The self-similarities found in the images are of three types. First type is exact self-similarity, second is quasi self-similarity and third is statistical self-similarity. This paper makes use of statistical self-similarity to achieve image compression. Unlike the traditional approaches we do not need two partitions of same image in the form of range blocks and domain blocks; instead a single partition of image is sufficient. This approach divides the code book into two disjoint subsets. One subset contains the blocks in which the pixel values are not much different and satisfying some threshold limit. This subset is the target area for compression. The statistical self-similarity based on the mean value is found among various blocks of this subset. The other subset contains the blocks that do not satisfy the threshold limit and therefore cannot be compressed. The overhead of image partition is halved in this approach; moreover since the target area of compression is separated the procedure is faster than the traditional approaches.
Keywords: Fractal Image Compression, Exact Self-similarity, Quasi Self-similarity, Statistical Self-similarity, Range Blocks and Domain Blocks
Download Full-Text
ABOUT THE AUTHORS
Jeet Kumar
*Qualification: M.Tech., M. Phil., MCA *Assistant Professor in Department of Computer Applications, *9 years experience in the field Academics and Research *More than 10 publications
Manish Kumar
Department of Computer Application, Shri Ramswaroop Memorial Group of Professional Colleges, Lucknow, Uttar Pradesh, India
Jeet Kumar
*Qualification: M.Tech., M. Phil., MCA *Assistant Professor in Department of Computer Applications, *9 years experience in the field Academics and Research *More than 10 publications
Manish Kumar
Department of Computer Application, Shri Ramswaroop Memorial Group of Professional Colleges, Lucknow, Uttar Pradesh, India
