MOMENT AND DENSITY BASED HADWRITTEN MARATHI NUMERAL RECOGNITION

In this paper, we present an OCR for handwritten Marathi Numerals recognition using two sets of features, namely, density and central moments. Central moments of order 3 and 4 are computed globallyfor the numeral image while density features are computed by dividing the numeral image into different zones. The combined features are fed to the classifiers namely, minimum distance, k-nearest neighbor (k-NN) and Support Vector Machines (SVM) to classify the unknown numeral patterns. The proposed method provided encouraging results.