Enhanced Fractal Image Compression using Quadtree Partitioning and Double Moment Descriptors

Abstract

Fractal Image Compression (FIC) is a robust method for reducing multimedia applications' data storage requirements and communication costs. This research introduces a partitioning strategy based on quadtree as variable length block partitioning with Prewitt Operator as a criterion for guiding the partitioning decision. The proposed system achieves significant time savings while increasing the Compression Ratio and identifies a near-optimal PSNR compared with the previous studies. The increase in PSNR reached 25% at its highest value while maintaining the compression ratio, which increased to 30% without a significant change in the time required for implementation. Additionally, a novel set of moment descriptors is introduced that developed in integration with the last set to highlight specific block regions, enhancing overall computational efficiency. The system operates on color images with different image sizes 512 × 512 pixels and 256 × 256 pixels, and color depth is 24 bits. This was also compared to verify the superiority of the results by the image size in pixels.  Experimental findings validate the efficacy of the enhanced FIC algorithm, demonstrating enhanced encoding speeds and improved image quality in the reconstructed output.