Handwritten English character recognition using convolutional neural network

Abstract

Converting paper material into electronic material is still a necessary work nowadays, however, recognition of handwritten characters still has limitations in their recognition rate, owing to the presence of various shapes, scales, and formats in different peoples handwritten characters. Machine learning has significant value in reducing human power. A Convolutional Neural Network model that is revised from LeNet-5, is used for handwritten letter recognition. This study uses the EMINST dataset to train the model, and the final recognition rate is about 93.44%.