Fuzzy acceptors for syntactic pattern recognition

Abstract The behavior of finite fuzzy acceptors over a bounded chain L is considered. The approximate (ϵ-) equivalence and ϵ-reduction by inputs are investigated. In the class of all finite fuzzy acceptors over L , the problems of ϵ-equivalence and ϵ-reduction are decidable. These results are implemented for pattern recognition of deformed images, as well as for restoration and enhancement of distorted patterns if the measure of the similarity is expressed by ϵ.