Compression of silhouette-like images based on WFA

Abstract

Summary form only given. The authors present the design a lossy fractal compression method for silhouette-like bi-level images that has an excellent quality to compression rate ratio. Their approach is based on weighted finite automata (WFA). We reduce the problem of the encoding of a silhouette-like bi-level image to the encoding of two one-variable functions describing the boundary (-ies) of the black and white regions of the given image. One advantage is that the automata encoding different bitplanes can share states.