Arabic letters recognition based on BP and perceptron and compare the performance of them

In this research we used a new method to distinguish the Arabic letter, depending on the different sizes of the Arabic letter and apply to all Arabic letters similar and dissimilar and took the character movements (opening, Vibrio, and Kasra) and will reach a good results.proposed algorithm working begin by introduce the Arabic letters, and a preliminary treatment in the image of the Arabic letters .in this research will be studied (matrix of a co-occurrences and characteristics of the matrix) as a parameter to distinguish the Arabic letter and give strong rate of recognition, applying network (perceptron) and a network (Back probagation) in order to speed up the process of recognition and make comparison between performance of the two networks.