Abstract:

Handwritten character recognition has gained significant attention in recent years, yet Arabic handwritten character recognition remains a challenging task due to the cursive nature of the script and variations in individual writing styles. Traditional machine learning approaches require handcrafted feature extraction and often fail to achieve high accuracy. To overcome these limitations, this study proposes a deep learning model based on Convolutional Neural Networks (CNNs) for Arabic handwritten character recognition. The model was trained and evaluated on the Arabic Handwritten Character Dataset (AHCD), which contains 16,800 characters written by 60 participants. Several preprocessing steps, including normalization, resizing, and data segmentation, were applied to improve input quality. The CNN architecture integrates batch normalization and dropout layers to prevent overfitting and enhance generalization. Experimental results demonstrate that the proposed model achieves a recognition accuracy of 98.39%, outperforming many traditional approaches and recent deep learning-based methods. These findings highlight the effectiveness of CNNs in recognizing complex Arabic handwritten characters and contribute to advancing robust solutions for Arabic Optical Character Recognition (OCR) systems.