Arabic Character recognition problem is one of the important issues that receives a lot of attention due to its great benefit in real-life applications like automates paperwork, mail and e-books, thus saving time and effort, but due to the low accuracy of discrimination, this issue is not applied in practice yet.
This article presents a high recognition accuracy for the printed Arabic character using deep learning based on the neuronal convolutional network, where the structure of these complex networks plays an important role in the distinctive success of the character , so we will conduct experiments on different network structures to study the possibility of success in each structure.
Al-Baath University Journal.
2020.
Detection of the Arabic letters using convolutional neural networks