Principal Component Classi cation for Fractal Volume Compression

Fractal volume compression is an extension of fractal image compression to volumetric data. The technique partitions the volume into range blocks which are replaced in the output dataset by transformations that map larger portions of the volume into blocks that resemble the original range partitions. Principal component analysis provides a classi cation system which directs the search for these self-similar maps to likely candidates. Classi cation also allows for simpler maps to represent less complex range blocks thus providing a higher compression rate.