Efficient Algorithm for Handwritten Digit Recognition based on Deep Learning

Abstract

Facial recognition in the field of image processing, which is a very important field in deep learning. In this work, the classification of numbers led to the training of the deep neural network and the Convolutional Neural Network (CNN) using the MATLAB program because of the large data carried by the color image, where a place was prepared to store information. create an advanced neural network for image recognition with the help of convolutional neural network training, which is the primary tool for deep learning. A network consisting of numbers from 0 to 9 was created, each number is a repository of 10,000 image data i.e. a table 10 x 10 so that the total number of images is 10,000 images stored and through deep learning the convolutional neural network was created Good results were obtained that proved the efficiency of the algorithm Proposed.