compression fractale

CONT: Le principe de la compression fractale réside dans la recherche de formes ou de motifs répétitifs dans une image, et dans le codage mathématique sous forme d'équations de ces motifs. Il est également possible de définir et d'exploiter des bibliothèques de motifs, afin d'améliorer la compression sur des séries d'images identiques.

CONT: Video compression ratios available from various vendors of full-motion video boards range up to 160:1 and perform adequately on standalone multimedia desktop platforms transmitting image data from CD-ROM or videodisk peripherals. However, these compression ratios will not be sufficient for effective transmission of complex videos and renderings that are being produced by engineering designers or even simpler videos in highly interactive multiuser networked environments. The solution, however, is already on the horizon in the form of fractional dimensional compression techniques (fractals) that can reach astonishing compression ratios of 10,000:1. This is possible because fractals, despite their apparent complexity, have a very low information content and can be specified as relatively simple expressions. Fractal compression is obtained via mathematical functions that describe a sequence of transformations of an image and when applied repeatedly, converge on a unique fractal. Fractal compression algorithms can be incorporated in software or VLSI microchips and, if applied to a typical CD-ROM disk, should provide nearly seven hours of full-motion.