Recognition of Handwritten Devanagari Words Using Neural Network

Handwritten Word Recognition is an important problem of Pattern Recognition. Online handwritten recognition system for Devanagari words is still in developing stage and becoming challenging due to the large complexity involvement. In India, more than 300 million people use Devanagari script for documentation. There has been a significant improvement in the research related to the recognition of printed as well as handwritten Devanagari text in the past few years. Though Devanagari is the script for Hindi, which is the official language of India, its character and word recognition pose great challenges due to large variety of symbols and their proximity in appearance. We have developed an offline recognition system to recognize the handwritten Devanagari Legal Amount words using neural network. First we have applied preprocessing techniques on scanned image, such as grayscale, binarization, and thinning techniques. For performing thinning, we used Stentiford algorithm. Then we have performed segmentation on thinned image. Finally we used recurrent neural network classifier as a recognition method. We have used database containing 26,720 handwritten Devanagari legal amount words written in Hindi and Marathi languages i.e. (Devanagari Script) by different writers. Out of these, we have trained 500 words as a one dataset for our proposed system. Result graph was plot, these training dataset against recognized word by the system.