Tifinagh Handwritten Character Recognition Using the Convolutional Neural Network
Sliman Rajaa, Azouaoui Ahmed · 2024
The Tifinagh alphabets have become a large field of research which attracts researchers due to the variety and similarity of these characters, in this article we abord a new recognition system aims to convert the Amazigh handwritten character input images to a machine-readable text based on the convolutional neural network, it consist of extracting the features from each image using convolution layers and then a neural network to train the classifier on these features in order to find a very good results and a high performance accuracy. The proposed optical character recognition is tested using a dataset handwritten character database and approves that achieved a good recognition accuracy (98.70%) comparing it with the previous work.