Computer vision and the creation of a database of printers’ ornaments

Research output: Contribution to journalArticlepeer-review


Colleges, School and Institutes


This article describes the creation of a database of over 1 million images of eighteenth-century printers’ ornaments (or fleurons) using computer vision, and how the database was refined using machine learning. The successes and limitations of the method used are discussed, and the purpose of the database for research in the humanities is outlined. The article concludes with a summary of recent developments in the project, which include the addition of a visual search function provided by the Seebibyte Project.


Original languageEnglish
Number of pages18
JournalDigital Humanities Quarterly
Issue number1
Publication statusPublished - 21 May 2021


  • computer vision, typography, Eighteenth Century Studies