Recognition Of Handwritten English Character Using Convolutional Neural Network

Abstract

In the domain of computer vision and image processing, one of the most active and difficult study fields is handwritten character recognition. It may be used as a reading tool for bank checks, for identifying characters on forms, and for a slew of other purposes. The optical character recognition of the papers is similar to documents produced by hand by a human. This OCR is put to use to improve the simplification of the process of character translation, which may be obtained from a broad range of file types, such as image and word document files. Researchers have made tremendous progress in HCR by making use of vast amounts of raw data and new breakthroughs in Deep Learning and Machine Learning algorithms. The fundamental purpose of this research paper is to give a solution for several techniques of handwriting recognition. These methods include the usage of touch input through a mobile screen as well as the use of an image file. CNN is used to identify characters in a test dataset in this work. Work on CNNs' capacity to detect characters from a picture dataset and their accuracy of recognition will be examined. Characters are recognized by CNN by comparing and contrasting their shapes and distinguishing characteristics. The dataset A_Z Handwritten was used to test our CNN implementation's handwriting accuracy and model gives the 100% result to recognize the character.