The Combination of Fuzzy Logic and Expert System for Arabic Character Recognition

In this paper fuzzy logic (FL) and expert system (ES) theories are studied with regard to their contribution to solving the problem of OCR (optical character recognition). These theories have improved the learning and adaptation capacities related to varying shapes where information is qualitative, inaccurate or incomplete. The use of these technologies FL and ES proves interesting, efficient, and necessary to recognize all Arabic character. This combination is very useful to improve the powerful of hybrid intelligent systems HIS in the field of OCR. The primary goal of this combination (FL, ES) is to classify and to recognize all presented unknown shapes. These theories must achieve these tasks: to classify characters, and to make ones way of intelligent recognition by ES_FL system capturing the behaviour of a human expert knowledge. The training has used 280 descended pictures of the database of ACR (Arabic character recognition). The results gotten of ACR databases are promising