Utilizing Adaptive Finite Automaton (AFA) to implement Adaptive Digitized Straight Line Segments (ADSLS) actuating as exploration automaton of a boundary, we propose an alternative for the available researches on dominant point detection in which primitives are composed by ADSLS. Consequently, this method is shown by simulations to be effective to represent adaptive regions of support and adequate for the complexities of real world scenarios like a shape classifier. Furthermore, even being based in the simple underlying mechanism of Finite Automaton (FA), ADSLS is able to adapt, reacting to circumstance stimuli in a single pass, also presenting learning capability.
Keywords: Adaptive Systems, Automata, Computational Geometry, Error Correction, Pattern Recognition
