A Novel Method for Handwritten Digit Recognition System

Abstract

Character recognition is crucial in the contemporary world. It can resolve more difficult issues while also simplifying human tasks. Character identification from handwriting is one illustration. This system is used extensively throughout the globe to identify zip codes or postal codes for mail sorting. Handwritten symbols can be recognized using a variety of methods. In this paper, two methods—pattern recognition and convolutional neural networks—are studied. (CNN). Both techniques are described, and various implementations of each strategy are also covered. Methods for pattern recognition include Bayesian Decision Theory, Nearest Neighbor Rule, and Linear Classification or Discrimination. Neural networks are used for shape identification, Chinese character recognition, and handwritten digit recognition. Use of neural networks for training and identification Key Word: Digit Recognition, Handwritten digits recognition, Convolutional Neural Network.