Bangla Handwritten Character Recognition Using Extended Convolutional Neural Network

The necessity of recognizing handwritten characters is increasing day by because its various applications. objective this paper to provide a sophisticated, effective and efficient way recognize classify Bangla characters. Here an extended convolutional neural network (CNN) model has been proposed Our CNN tested on “BanglalLekha-Isolated” dataset where there are 10 classes for digits, 11 vowels 39 consonants. shows accuracy recognition as: 99.50% 93.18% vowels, 90.00% consonants 92.25% combined classes.