Image compression method based on Generalized Finite Automata

In this paper, we introduce an approach to compress gray image using deterministic generalized finite automata (GFA). By detecting the self-similarity inside an input digitized gray image, a GFA can be constructed to describe the image. The decode algorithm can restore the image from the deterministic generalized finite automata efficiently. This method has a smaller number of states than an equivalent classical finite automaton. Meanwhile it also has an advantage of higher compression without further degradation of quality.