Handwritten Recognition of Character and Number Using Convolutional Neural Network and Support Vector Machine

Abstract

Handwritten character recognition has become a major study topic as a result of the rising use of digital technology in many industries and in practically all day-to-day activities to store and convey information. Although handwritten copies are still useful, individuals prefer to have them turned into electronic versions that can be shared and saved online. Handwritten recognition is the capacity of a computer to recognize and understand comprehensible handwritten input from a variety of sources, including touch screens, pictures, paper documents, and other sources. Because diverse people have distinct handwriting styles, handwritten characters remain complicated. The purpose of this work is to describe the creation of a handwritten character and number recognition system that will be used to read handwritten notes from students and lecturers. For feature extraction and higher end classification, the Learning model uses Convolution Neural Networks (CNN) and Support Vector Machines (SVM). Handwritten inputs are scanned, noise is removed, and numbers and characters are retrieved to complete this operation.