Generalized Weighted Finite Automata Based Image Compression
Karel Culik II. (Department of Computer Science University of South Carolina Columbia, USA)
Peter C. von Rosenberg (Department of Computer Science University of South Carolina Columbia, USA)
Abstract: The Culik-Kari recursive inference algorithm for WFA is based on an effcient way of expressing subsquares of the given image as linear combinations of available states. Here we improve it in two ways. First, we allow the use of rotations, flippings and negations of the states in the linear combination. Second, in order to get the best possible representation of simple fractal images we allow the creation of edges pointing to ancestors of states under construction which, for technical reasons, was not done in the original recursive algorithm.
Keywords: WFA, fractal-image compression, image-data compression, nite automata