Recognition of Handwritten Persian/Arabic Numerals Based on Robust Feature Set and K-NN Classifier

Persian handwritten numerals recognition has been a frontier area of research for the last few decades under pattern recognition. Recognition of handwritten numerals is a difficult task owing to various writing styles of individuals. A robust and efficient method for Persian/Arabic handwritten numerals recognition based on K Nearest Neighbors (K-NN) classifier is presented in this paper. The system first prepares a contour form of the handwritten numerals, then the transit, angle and distance features information about the character is extracted and finally K-NN classifier is used to character recognition. Angle, transit and distance features of a character have been computed based on distribution of points on the bitmap image of character. In K-NN method, the Euclidean distance between testing point and reference points is calculated in order to find the k-nearest neighbors. We evaluated our method on 20,000 handwritten samples of Persian numerals. Using 15,000 samples for training, we tested our method on other 5,000 samples and obtained 99.82% correct recognition rate. Further, we obtained 89.90% accuracy using four-fold cross validation technique on 20,000 dataset.