Statistical analysis of objective measures using Fractal Image Compression for medical images

Abstract

Image compression helps in improving the transmission speed and better utilization of the Bandwidth by providing enough space for storing voluminous data. Fractal Image Compression (FIC), is one such lossy compression scheme based on contractive mapping theorem. FIC adopts affine transforms to map the range blocks and domain blocks by using the property of self similarity in the images. AS there is a need to enhance the compression performance in terms of compression ratio (CR) and peak signal to noise ratio (PSNR), an attempt is made in the present paper, to analyze FIC for Objective Parametric variations using Quad-tree Partitioning technique for Medical Image. In this work Medical Images like mammograms and MR Images of Brain are chosen in large numbers and analysis using quad-FIC with and without using progressive encoding technique is implemented. The Performance measures like Compression Ratio (CR), Peak signal to noise ratio (PSNR) for different Threshold Values are measures. Mean and standard deviations for PSNR and CR values, for a large set of data of one class of Image and a small set of data for different class of image was measured. A set of 100 mammogram images and 50 MR Images were considered for the work. Mat lab simulated results shows that the average standard deviations for PSNR value for large data and small data is less and the CR is uniform for throughout. The PSNR value is more for Mammogram images. Hence Mammograms are best suited for the Implementation with FIC. High Compression Ratio of 19.9 is achieved when FIC is implemented with Embedded Block coding optimization truncation (EBCOT) — a progressive encoding technique.