Handwritten character recognition models based on convolutional neural networks

Abstract

This paper explores the possibility of using various convolutional neural network models to recognize handwritten Arabic characters. It discusses which of the existing convolutional neural network models best suits this task. The results are analyzed based on which conclusions are drawn regarding the effectiveness or ineffectiveness of specific convolutional neural network models. The paper provides an analysis of the efficiency of convolutional neural network models for the task of recognizing handwritten Arabic characters.