A Systematic Literature Review of Deep Learning Methods for Handwritten Text Recognition in Historical Arabic Manuscripts
Bilal A. Tuama, Farhan Mohamed · Engineering Technology & Applied Science Research · 2025
Arabic Handwritten Text Recognition (AHTR) in historical manuscripts poses significant challenges due to the script’s cursive nature, variability in calligraphy styles, and document degradation over time. This paper presents a Systematic Literature Review (SLR) of recent Deep Learning (DL)-based approaches applied to AHTR, focusing on methods developed between 2020 and 2025. It analyzes key DL architectures and provide a comparative analysis of the most commonly used datasets and segmentation strategies. Additionally, this review highlights essential preprocessing and postprocessing techniques that enhance the recognition performance and discusses the common evaluation metrics for AHTR. Finally, it identifies the current challenges and proposes future research directions to improve the recognition accuracy and model generalization. This review aims to guide researchers in building more robust and effective systems for the preservation and digitization of Arabic cultural heritage.