Identity Card Number Recognition System using Artificial Neural Network

Abstract

This paper proposes a system for recognizing the identity card number. The quality of the color image of identity card is enhanced by image filtering technique. Color sensing is carried out to get the image portion of the identity card number . Noise canceling is performed to remove the unwanted pixel areas. The characters and digits are segmented out individually by horizontal and vertical Histogram projections. The individual images of printed Myanmar characters and handwritten Myanmar digits are resized into predetermined size to get uniformity. These preprocessed images are applied into the ANN of the proposed system. The ANN is de signed with three layers; input , hidden and output layer. The output values of the output layer are the certaint y values of digits and characters. The highest certainty output value is the digit or character that is recognized by the network. After all the digits and characters are recognized, the proposed system identifies out the identity card number.