Handwriting Arabic character recognition Lenet using neural network

Abstract EN

Character recognition has served as one of the principal proving grounds for neural network methods and has emerged as one of the most successful applications of this technology.

In this paper, a new network is designed to recognize a set of handwritten arabic characters.

This new network consists of two stages.

The first is to recognize the main shape of the character, and the second stage is for dots recognition.

Also, the characteristics, structure, and the training algorithm for the network are presented.