Southeast Asian Illuminated Manuscripts: A Study of the Maritime Silk Road Using Scientific Imaging and AI

Lead Research Organisation: Nottingham Trent University
Department Name: Graduate School


This project would investigate how machine learning or AI can be used in conjunction with noninvasive
imaging techniques to determine multiple characteristics of maritime Southeast Asian
manuscripts. AI analytical methods would be designed and developed so that detailed information can
be found to answer many fundamental questions. To acquire the data aiding this task, imaging
methods such as X-ray fluorescence, Spectral Imaging, and OCT would be used so that multimodal
analysis could be performed using AI automation. My relevant experience includes an MSci Research
Project involving in-depth spectral analysis of two Southeast Asian Islamic Manuscripts belonging to
the British Library. The manuscripts in question contained colophons stating their origins of writing in
locations within Southeast Asia, however the hand of the script, illumination and other aspects of the
manuscripts suggested they instead had been written in Dagestan, Russia. I performed spectral
comparisons between coloured regions that were seemingly the same, however, the spectra recorded
were not, suggesting that the colophon had been added later. When doing the analysis to come to
these conclusions I developed my own programmes to calibrate and extract important data
automatically, allowing for quicker analysis of large areas. I have previously studied English to AS level
and have a keen interest in historical literature. I also recently partook in a field trip to the Abbey Library
of St. Gall in Switzerland where I performed Multi-Spectral, Ultraviolet Fluorescence, and Near Infra-red
imaging on centuries old manuscripts of great historical and religious importance. The fieldwork
allowed me to become accustomed to the importance of such manuscripts and provided large
amounts of hands on experience with imaging equipment and methods. During this I also performed
preliminary analysis of two manuscripts using automated techniques to uncover hidden writing and


10 25 50