Ancient script AI reads and translates damaged inscriptions automatically
InscriptionOCR: A Dataset and Method for Understanding Inscriptions
Computer Vision and Pattern Recognition
Summary
Ancient writings carved in stone are hard to read because the images are often damaged and there are few digital tools available. The authors created an AI system that cleans up the images, recognizes ancient Brahmi characters, converts them into the familiar Roman alphabet, and then translates the old Prakrit language into English. They also made large digital datasets with many character images and sentence translations to help train and test these tools. This work helps people study old inscriptions more easily using computers.
What this means in practice
- •For cultural heritage teams: Automatically generate readable and translated text from photos of ancient stone inscriptions using AI processing.
- •For language technology developers: Build OCR and translation tools for scripts and languages that have limited digital resources using the provided datasets and methods.
Authors
Jaidev Sanjay Khalane, Akbar Ali, V. N. Prabhakar, Shanmuganathan Raman
Abstract
Ancient script image restoration is a fundamental problem in computer vision, as it directly affects the reliable analysis and interpretation of historical documents and inscriptions. Ashokan Brahmi is an ancient script extensively used during the reign of Emperor Ashoka in the 3rd century BC, primarily for inscriptions in Prakrit. These inscriptions, including major and minor rock and pillar edicts, constitute a valuable yet largely unexplored source of data for computational analysis. The degraded nature of inscription imagery and the lack of standardized digital resources pose significant challenges for automated processing. We present an end-to-end AI-based framework for understanding ancient inscriptions that encompasses image enhancement, optical character recognition (OCR), transliteration, and neural machine translation (NMT). The proposed pipeline processes low-quality images captured directly from stone inscriptions, performs image restoration and Brahmi script character recognition, maps the recognized characters to the Roman script, and finally translates the resulting Prakrit text into English. We also introduce two new datasets: (i) InscriptionOCR Dataset: the largest publicly usable digital OCR dataset for Brahmi script to date, consisting of over 200,000 character images across about 600 classes, and (ii) a bilingual Prakrit-English parallel corpus comprising over 2,000 sentence pairs for NMT. We believe that the proposed framework and datasets will facilitate future research in ancient script analysis, low-resource OCR, and digital epigraphy.