Adaptive Fractal Coding of Still Pictures

This paper presents a method for lossy coding of still pictures using the theory of iterated function systems (IFS) which is a standard description for deterministic fractals. The coding concept is based on a blockwise fractal approximation of the original image by contraction mappings of itself using affine transformations which are a special case of IFSs. Data reduction is achieved by exploiting the self-similarity within natural or artificial pictures. The proposed algorithm accounts for the local image characteristics by an adaptive quadtree segmentation. In contrast to existing fractal coding concepts an overall constant reconstruction quality is achieved.