期刊名称：International Journal of Advances in Engineering and Management
电子版ISSN：2395-5252
出版年度：2022
卷号：4
期号：4
页码：1053-1060
DOI：10.35629/5252-0404824828
语种：English
出版社：IJAEM JOURNAL
摘要：Handwritten Character Recognition is a subfield of Image Processing that deals with extracting texts from images or scanned documents.Because of its wide range of applications, such as automatic bank check processing, billing, and automatic postal service, handwritten digit recognition has a large research area.It is difficult for a machine to comprehend handwritten numerals because of the variations in shape and orientation.In this paper, Convolution Neural Network (CNN) was employed to recognize isolated handwritten digits (0 to 9). We have chosen to take the handwritten images for both training and testing from the MNIST dataset. The handwritten recognition system allows users to draw a digit on a webpage, renders it into an image and then recognizes the features of the image and classifies it into one of the ten predefined digits (from 0-9). It thendisplays a performance curve based on the acquired accuracy. The main objective is to recognize the handwritten digits with maximum accuracy.
