A fast edge preserving fractal system

Scope and method of study. In this research, we present a fast edge preserving fractal system, FEPFS. For applications, such as satellite surveillance, fractal block coding provides a method to compress digital images at high compression ratios. Incorporation of an edge based quad-tree partitioning and a wavelet r-tree search engine allow fractal block coding to be used in almost 'real time'. Diffusion techniques enhance and restore edge information in the attractor of fractal code creating a more robust method of image compression.;Findings and conclusions. The fast edge preserving fractal system, FEPFS, provides a practical compression methodology for near 'real time' application like surveillance. Based on the correlation of self-similarity, fractal coders compress a digital image by relating pairs of image blocks at different scales in the image. Searching for 'optimal' pairs constitutes the major computation load for fractal coders. Implementation of the wavelet r-tree search engine reduces the computational complexity, search time, and memory requirements. The r-tree search engine reduces the search time to seconds by providing a dynamic index structure for spatial searching of images. Inaccurate mappings at large block sizes limit the compression ratios of fractal coders. The attractor of the fractal code suffers from edge degradation and spurious artifacts at large block sizes. FEPFS uses diffusion techniques to preserve significant edge information at a lower bit rate cost than partitioning down to small block sizes. By expanding the basic diffusion equation to contain a scalar function based on the edges of the original image, FEPFS directs the diffusion process to smooth along the direction of significant edges and sharpen in the direction of the edges. Using this method of diffusion, FEPFS restores edge information and smooth discontinuities and blocking effects.