The system to recognize the Chinese characters written in a firm hand isdesigned. With the idea that the adaptive observations play the important rolein recognizing complicated patterns, this paper shows one of the attempts tomake a machine have such a function.The system proposed here can be characterized with two points of new approaches.The first is a hierarchical structure that is introduced into the representationof Chinese characters. The second is an adaptive function of therecognizer, which makes possible to analyze and recognize Chinese charactersspoiled in a topological sense.Owing to the adaptive function realized by employing the concept of aprovisional recognition, the most adequate manner of observations for eachobject can be taken at the stage of measurement. The details ot the precedingcharacteristics and the concrete description of the system possessing thesecharacteristics are shown. Finally, the ability of the proposed system is discussedon the whole.
展开▼
