Efficient Recognition of Devanagari Handwritten Text

Research in Devanagari handwritten character recognition has challenging issues due to the complex structural properties of the script. Handwritten characters are of unique style for individual persons. The process of recognition involves the preprocessing and segmentation of input document image containing Devanagari characters. It is proposed to use Kohonen Neural Network to interpret Devanagari Characters from the segmented images. The recognized Devanagari characters are stored in a file.