Recognition of Handwritten Devanagari Character using Convolutional Neural Network
Indu Dokare, Siddhesh Gadge, Kedar Kharde, Siddhesh Bhere, Rohit Jadhav · 2021
The use of a Convolutional Neural Network to recognize Devanagari characters is explored in this paper. Any pattern recognition task’s classification stages and feature extraction are responsible for accurately defining the patterns. Deep learning surmounts the part of highlight extraction and does so consequently, encouraging the weight on software engineers. Deep learning is slowly but steadily displacing other pattern recognition methods. Deep learning is the best option to tackle the complexities involved in applications like character recognition, which require vast quantities of data and uncertainty in the data. Devanagari consonants have a recognition accuracy of 98 %, vowels have a recognition accuracy of 97.56 %, and Devanagari numerals have a recognition accuracy of 99 %.