Fractal Encoding by Classified Domain Trees

Fractal coding of digital images o ers many promising qualities. The encoding process, however, su ers from the long search time of the domain block pool. A standard technique for speeding up the encoding is the feature classi cation of blocks. In this paper we show that the classes can be arranged in a tree, and searched by a pruning algorithm developed and explained in a previous paper by the author. We show that this hybrid method enhances the search speed by upto 75% over the purely feature based method.