Low complexity fractal-based image compression technique

Abstract

A fast image compression technique as well as its progressive image transmission (PIT) version using fractals is presented which uses a small pool of domains extracted using visually significant patterns. The affine transformations for an edge block are obtained by using its edge characteristics instead of the minimum mean square error criterion. Simulation studies demonstrate that this method is computationally simple, gives faster encoding speed and achieves good fidelity at relatively higher compression ratios than other fractal based techniques.