DEVNAGRI SCRIPT RECOGNITION USING ARTIFICIAL NEURAL NETWORK CLASSIFIER

Kasturi Upasani, P. V. Baviskar

Abstract: In India, Devnagari script is used by 500 million people and more. Now-a-days Special attention is given to Devnagari so that examination or investigation of ancient Indian literature can be done efficiently. The paper focus on the innovative methods for stepwise segmentation line segmentation and segmentation of character which are overlapping characters of Devnagri text. The steps followed to complete the process initially the character are segmented into lines, lines into words and then from each line the header lines of words are detected and then transformed as straight lines. Finally artificial neural network is used for classification.

Keywords: Devnagari, Backpropagation, and Blob Analysis.

DOI: https://doi.org/10.15623/ijret.2016.0506051