AI-Assisted Digitization and Text Reconstruction of Ancient Manuscripts: Revitalizing Tarjuman for the Digital Era
The Islamic intellectual history of Madura and East Java is deeply anchored by the works of local ulema who translated complex sacred knowledge into vernacular understanding. Among the most influential figures is Kiai Haji Abdul Hamid bin Isbat, a prominent scholar whose classical work, Kitab Tarjuman, served as an essential bridge for local students (santri) to comprehend Islamic teachings, Quranic exegesis, and religious guidance.
Written primarily using the Pegon script—an adaptation of Arabic letters used to write vernacular languages such as Madurese and Javanese—Kitab Tarjuman captures both theological depth and local linguistic heritage.
The Core Problem: Physical copies of Kitab Tarjuman are vulnerable to aging paper, fading ink, and limited physical distribution, risking the loss of KH Abdul Hamid bin Isbat's intellectual contributions.
By integrating Artificial Intelligence (AI) with classical philology, researchers can digitize, restore, and analyze Kitab Tarjuman, transforming a fragile historical manuscript into an accessible, searchable digital resource.
Technological Reconstruction of Pegon Manuscripts
Deciphering historical manuscripts written by hand presents unique technical challenges. Kitab Tarjuman utilizes Pegon script, which includes specific diacritics and character variants designed for Madurese phonetics that standard digital tools fail to recognize.
AI addresses these challenges through a three-stage preservation pipeline:
Visual Enhancement & Artifact Cleaning
Physical degradation such as ink bleed-through, paper discoloration, and torn margins obscures Kyai Hamid's original writing.
Generative Adversarial Networks (GANs): AI isolates background stains and removes ink bleeding from reverse pages.
Contrast Optimization: Enhances stroke clarity without altering the original handwritten calligraphic style.
Specialized Handwritten Text Recognition (HTR)
Standard OCR models cannot process non-standardized Pegon handwriting.
Custom Pegon Datasets: HTR neural networks are trained specifically on Pegon orthography and KH Abdul Hamid bin Isbat’s distinctive writing style.
Character Mapping: Converts physical handwriting into standardized digital Unicode text.
Contextual Inpainting & Gap Reconstruction
Where pages are damaged or eaten by insects, domain-adapted Large Language Models (LLMs) analyze surrounding sentences to propose plausible text restorations based on classical Pegon grammar and traditional pesantren references.
Transliteration, Translation, and Religious Analysis
Once Kitab Tarjuman is digitized into machine-readable text, AI algorithms enable deep linguistic and thematic exploration of Kyai Hamid's scholarship.
Automatic Transliteration
Fewer young scholars today read fluent Pegon. The AI system automatically converts Kyai Hamid’s Pegon text into modern Romanized script (Rumi), making Kitab Tarjuman instantly readable for broader audiences.
Thematic & Semantic Mapping
Topic Categorization: Automatically groups passages into Islamic jurisprudence (fiqh), ethics (tasawwuf), and creed (aqidah).
Cross-Manuscript Alignment: Compares different surviving physical copies (naskhah) of Kitab Tarjuman preserved across various pesantren archives to identify copyist variations (scribal errors).
Impact on Pesantren Heritage and Digital Humanities
Applying AI to KH Abdul Hamid bin Isbat's work demonstrates how modern technology can revitalize local Nusantara scholarship rather than replace traditional learning.
Traditional Philology vs. AI-Assisted Workflow
While traditional philology relies on painstaking manual scholarship to preserve context, the AI-assisted pipeline introduces an unprecedented leap in efficiency and global accessibility. Automating transcription and cross-text alignment drastically reduces processing times from months to minutes, while automated Pegon-to-Rumi transliteration removes linguistic barriers for non-specialist readers.
The primary trade-off involves shifting initial effort away from manual deciphering toward training robust neural models, verifying AI-generated output for domain accuracy, and ensuring high-fidelity digitizations. Ultimately, the AI-assisted workflow does not replace expert scholarship; rather, it complements traditional methods by safeguarding fragile manuscripts against physical decay in permanent digital archives and freeing researchers to focus on high-level analytical and theological interpretation.
The AI-driven digitization of Kitab Tarjuman ensures that KH Abdul Hamid bin Isbat’s intellectual legacy remains active, searchable, and accessible. By bridging traditional pesantren manuscript heritage with modern digital humanities, this project safeguards a vital chapter of Islamic scholarship for future generations.
SDGs Addressed
 | Through the preservation of ancient Islamic texts meant for educating future generations |
 | Through development of an innovative solution of saving old and damaged manuscripts with the use of Artificial Intelligence (AI) |
Researcher Featured:
Dr. Imamuddin bin Nahrawi
Dean’s Office,
Academy of Islamic Studies, Universiti Malaya
Area of Expertise:
Quranic Literacy > Shariah Research > Islamic Education
Get to know the researcher:
Dr Imamuddin bin Nahrawi is a scholar of Shariah and Maqasid Shariah whose work focuses on making Islamic knowledge meaningful and relevant to contemporary society. Educated in Indonesia, Egypt, and Malaysia, he combines rigorous academic scholarship with extensive experience in community engagement, Quranic education, and public outreach.
For further inquiries, please contact:
Email: imamuddin@um.edu.my
Tel: +603-7967 6183
Web of Science (WoS): https://www.webofscience.com/wos/author/record/Imamuddin
Author:
Jaiyogesh Ramesh Patel
Research Outreach & Visibility Centre (OVIS), JPP, UM
Copy-editor:
Siti Farhana Bajunid Shakeeb Arsalaan Bajunid
Research Outreach & Visibility Centre (OVIS), JPP, UM
Comments