Abstract: This paper describes a method of Automatic Arabic Character Recognition (ACR)

This paper describes a method of Automatic Arabic Character Recognition (ACR). This system is articulated around four distinct modules which are the main image processing given by: a module of treatment, a module of segmentation, a module of recognition and a module of detection of symbols of classification. For classification; we have used Fuzzy Logic (FL), Genetic Algorithm (GA), and Expert System (ES) to recognize all presented shapes. This approach must make the system able to achieve these tasks: to extract the components of QV (Quantification Vector) and to take an intelligent decision to classify all different shapes of Arabic characters. The Fuzzy classifier is combined with the Expert System to extract the topological and contextual informations of each character. The exit of the system will be combined with the one of the recognition module for the reconstitution of each character. The training has used 180 descended pictures of the database of ACR. To deal with Artificial Intelligence (AI), a new design methodology has been introduced to give a best rate of recognition. The new design methodology based upon a VHDL description of recognition has the two (02) advantages: to present a real behaviour of good automatic system recognition for Arabic texts, and being generic and flexible and can be changed at the user demand. To validate our approach, a VHDL description of ES_GA_FL is passed through a synthetic tool, Galileo for FPGA implementation.