Statistical Analysis of Fractal Image Coding and Fixed Size Partitioning Scheme

Fractal Image Compression (FIC) is a state of the art technique used for high compression ratio. But it lacks behind in its encoding time requirements. In this method an image is divided into non-overlapping range blocks and overlapping domain blocks. The total number of domain blocks is larger than the range blocks. Similarly the sizes of the domain blocks are twice larger than the range blocks. Together all domain blocks creates a domain pool. A range block is compared with all possible domains block for similarity measure. So the domain is decimated for a proper domainrange comparison. In this paper a novel domain pool decimation and reduction technique has been developed which uses the median as a measure of the central tendency instead of the mean (or average) of the domain pixel values. However this process is very time consuming. Keyward: fractal image compression, fishers classification, hierarchi-cal classification, median, DCT, IFS, PIFS, PSNR. GJCST-F Classification: I.3.3 StatisticalAnalysisofFractalImageCodingandFixedSizePartitioningScheme Strictly as per the compliance and regulations of: Statistical Analysis of Fractal Image Coding and Fixed Size Partitioning Scheme AbstractFractal Image Compression (FIC) is a state of the art technique used for high compression ratio. But it lacks behind in its encoding time requirements. In this method an image is divided into non-overlapping range blocks and overlapping domain blocks. The total number of domain blocks is larger than the range blocks. Similarly the sizes of the domain blocks are twice larger than the range blocks. Together all domain blocks creates a domain pool. A range block is compared with all possible domains block for similarity measure. So the domain is decimated for a proper domain-range comparison. In this paper a novel domain pool decimation and reduction technique has been developed which uses the median as a measure of the central tendency instead of the mean (or average) of the domain pixel values. However this process is very time consuming. Thus another technique has been suggested which heuristically eliminates the empty domain classes. Experiments on some standard image data shows that the proposed technique improves the PSNR of the decompressed image when compared with baseline fractal image compression (BFIC) and comparable with other scheme proposed till date.