Abstract
Video encoding techniques play an important role in data reduction. Fractal compression has received considerable attention in the past decades. While early methods presented prohibitively large encoding times, recent search less fractal encoders reduced this problem. A fast 3D purely fractal video encoder based on a flexible adaptive spatial subdivision data structure is proposed in this work. The method completely avoids any kind of search for a matching domain block and is capable of performing fast compression and decompression with high visual fidelity. Experimental results show that the developed approach outperforms the state-of-the-art x264 video encoder at very low bit rates in high motion video sequences in both structural dissimilarity measure and encoding time.