Artificial Neural Network Application In Letters Recognition For FarsiArabic Manuscripts

Abstract- Letter recognition for manuscript is one of the categories that has been deliberated in recent years and has many applications. Considering variety of hand writings correct recognition of manuscript letter has many difficulties. In literature various algorithms has been used to letter recognition for manuscript in different languages. Regarding to artificial neural networks ANNs abilities in machine learning parallel processing flexibility and pattern recognition it would be a convenient method to be used in this field. In this paper we proposed an ANN based algorithm to letter recognition for FarsiArabic manuscript. Finally we illustrate that proposed method is one of the best method to be used in letter recognition for FarsiArabic manuscript.