Data set for Tifinagh handwriting character recognition

Area of Science:

• Linguistics
• Computer Science
• Artificial Intelligence

Background:

• The Tifinagh-IRCAM alphabet is the official script for the Amazigh language, prevalent in North Africa.
• Despite its official normalization in Unicode in 2003, the Tifinagh character repertoire requires further development.
• Existing resources for Tifinagh character recognition and analysis are limited.

Purpose of the Study:

• To address the need for more resources in Tifinagh character recognition.
• To introduce a comprehensive dataset for training and evaluating Tifinagh character recognition systems.
• To facilitate the extraction of meaningful characteristics from individual Tifinagh characters.

Main Methods:

• A dataset of 1376 handwritten Tifinagh character images was created.
• Each of the 33 Tifinagh characters (including labialized forms) is represented by 43 images.
• The dataset is designed for use in machine learning models for character recognition.

Main Results:

• A novel, high-quality dataset of handwritten Tifinagh characters is now available.
• The dataset provides a robust foundation for developing automated Tifinagh character recognition.
• The dataset enables detailed analysis of character features for linguistic research.

Conclusions:

• The proposed dataset is a significant contribution to Amazigh language processing and digital humanities.
• This resource is expected to accelerate research and development in Tifinagh character recognition.
• Further work on Tifinagh character datasets will enhance its utility and applications.