Computational Analysis of Digitized Images from the Roman de la Rose Digital Library

Association for Computers and the Humanities
Pittsburgh, PA
July 23-26, 2019

Proposal

The Roman de la Rose Digital Library (RDL) (https://dlmm.library.jhu.edu/en/romandelarose/) serves as a resource for the study of the most popular secular work of the European Middle Ages by providing access to 146 digitized manuscripts in IIIF format and with additional datasets describing the manuscripts in the corpus. I have been working with RDL data to create an interactive visualization platform for exploring codicological and location information. This next phase of the project employs computational image analysis to explore the digitized images themselves in conjunction with and in comparison to the codicological data provided by the RDL. Applying new methods to this corpus will open new avenues of research for medieval studies scholars interested in the history of the book, illustration transmission, and more. 

I will use Imageplot software (http://lab.softwarestudies.com/p/imageplot.html) to explore issues of image saturation and brightness. Such calculations will be put in comparison with the codicological dataset provided by the RDL. One would expect manuscripts with a higher median brightness to reflect manuscripts with larger borders and/or to have few to no illustrations, for example. Finding evidence to challenge or support codicological analysis will further research understandings of this well-studied corpus. 

The Distant Viewing Toolkit (https://www.distantviewing.org/) (DVT) will allow analysis of facial and object detection across the digitized manuscripts in the corpus that have at least one illustration. The RDL has datasets of Illustration Titles and Narrative Sections, providing insight into characters and their frequency across the many variations of the corpus. Bringing the RDL datasets into comparison with algorithmic analysis of characters may reveal new understandings of the consistency of figures in illustrations.

On the whole, this work seeks to bring a robust dataset of digitized manuscript images into contact with multiple image analysis approaches to test their use in medieval manuscript analysis. I hope to spark new conversations among medieval studies scholars, historians of the book, art historians, and scholars interested in computer vision and computational image analysis.

Leave a Reply

Your email address will not be published. Required fields are marked *