What type of study is this?

This is a Quantitative Study study.

October 8, 2025Open Access

Bridging the Gap: Fusing CNNs and Transformers to Decode the Elegance of Handwritten Arabic Script

Key Points

The ensemble model achieved a letter classification accuracy of 96.38% on the IFN/ENIT dataset, significantly improving recognition performance.
Results also indicated a positional classification accuracy of 97.22%, showcasing the hybrid model's effectiveness in handling dynamic letter forms.
This work combines CNN and Transformer architectures, utilizing their strengths to enhance the capabilities of optical character recognition systems.
The findings suggest that integrating CNNs and Transformers may advance real-world applications in handwriting recognition technologies.

Abstract

Handwritten Arabic script recognition is a challenging task due to the script's dynamic letter forms and contextual variations. This paper proposes a hybrid approach combining convolutional neural networks (CNNs) and Transformer-based architectures to address these complexities. We evaluated custom and fine-tuned models, including EfficientNet-B7 and Vision Transformer (ViT-B16), and introduced an ensemble model that leverages confidence-based fusion to integrate their strengths. Our ensemble achieves remarkable performance on the IFN/ENIT dataset, with 96.38% accuracy for letter classification and 97.22% for positional classification. The results highlight the complementary nature of CNNs and Transformers, demonstrating their combined potential for robust Arabic handwriting recognition. This work advances OCR systems, offering a scalable solution for real-world applications.

Read Full Paperexternally

AI에게 질문

Bookmark

View Full Paper