Recognition of Arabic Handwritten Amount in Cheque through Windowing Approach

Arabic language is a semantic language that has differences when compared to English language. We are dealing with the handwritten Arabic Amount from cheques of Arabic banks . In this paper we proposed a windowing technique for the segmentation of the numerical amount, followed by an efficient moment invariants for features extraction . A maximum and minimum points technique used to isolate the Arabic (Hindi Digits) numerals. The feature vectors are grouped for each digit and Artificial Neural Network (ANN), is applied for the classification and recognition. This approach resulted in providing 99.5% of recognition rate.