Handwritten Digit Recognition Using Computer Vision

Abstract

Recently, with the rise of the Artificial Neural Network (ANN), Deep Learning Machine has made a dramatic turn in the field of art by becoming more and more intelligent. Intensive learning is used in the field of intensive learning due to a variety of applications including intelligence, health, medicine, sports, robotics, and drones. In deep learning, the Conventional Neural Network (CNN) is at the heart of the heart with great goals. The Artificial Neural Network (ANN) and the latest deep light are widely used in pattern recognition, sentence classification, speech recognition, face recognition, text classification, document analysis, visualization, and handwritten digit recognition. This paper aims to classify handwritten digits using different numbers of hidden layers and ages and to observe the variability of the accuracy of CNN to make comparisons between accuracies. To evaluate this performance of CNN, we performed our experiment in the modified National Institute of Standards and Technology (MNIST) dataset. This paper distinguishes between those functions using some form of data augmentation and works out of the box with the original dataset. Also, functions using CNN are reported separately. This is a sophisticated yet reliable approach in solving the problem. Nowadays, considerable work has been achieved with a test error rate of less than 1% in this dataset