A RESEARCH ON IMPROVE HANDWRITTEN CHARACTER RECOGNITION BY USING CONVOLUTIONAL NEURAL NETWORK

Abstract

For image recognition CNN is the most popular learning model. The features like weight sharing strategy
and strong relations of the pixels of the image makes CNN best choice for image recognition. The feature extraction and
classification can be done simultaneously in deep learning models which has proved very needful compared to the
traditional methods. A promising recognition can be obtained by using CNN if we address to certain issues. So in CNN
based framework for handwritten character recognition that gives a better performance compared to other CNN based
recognition methods.