What type of study is this?

September 10, 2025Open Access

Digital Codicology on the Vatican Apostolic Library

Key Points

The analysis of 1,650 manuscripts shows significant shifts in codicological parameters over centuries.
Key findings include the development of eight codicological parameters, with data sourced from 600,000 images.
Digital codicological analysis involved integration of high-resolution images and machine learning techniques.
The findings highlight potential for automation in dating Western medieval manuscripts with neural network technology.

Abstract

1650 Western manuscripts written between the year 800 AD and the year 1500 AD have been subjected to digital codicological analysis. The analysis involved approximately 600,000 high-resolution images (400 dpi), enabling the automatic extraction of 8 codicological parameters: Filling coefficient, percentage of ink on the entire page, percentage of ink in the written area, number of rows, line spacing, text height, page size, and page ratio. This study shows the evolution of these 8 parameters over the centuries, and finally a first neural network capable of automatically dating Western medieval manuscripts has been built.

Read Full Paperexternally

Ask AI

Helpful

Bookmark

View Full Paper