IPOP: Image recognition in pre-modern, popular print collections

About the project

The aim of this project is to use computer vision techniques developed in the Clariah Media Suite to perform image recognition on a collection of Dutch (printed) penny prints. The massively produced, Dutch penny prints are an interesting use case for this Media suite, because of their strong visual appearance, but also because of their specific visual features (8 to 24 woodcuts, with short captions). Image recognition is a crucial additional research tool in order to study and compare the longevity, materiality, similarities, differences and exchanges of visual material in this popular Dutch genre.

We do know for instance that the woodblocks were used for a long period of time in the printings shops, sometimes even many decades. We also know that these woodblocks were often traded between booksellers and thus ended up in other editions and even other genres. When we are able to trace the use of these images and woodblocks via image recognition, we can reconstruct production networks (printers, publishers, engravers) and analyse the use and adaptation of these images in different contexts and periods. Furthermore, we can see how representations of certain events, scenes, objects and protagonists are changed and adapted over time.

Last but not least, we will eventually (when the Delpher collections are incorporated) be able to trace these images moving from one genre to another, and thereby increasing our insights in popularisation and adaptation patterns. For this specific fellowship we will select 150 penny prints from the Delpher collection ‘Het Geheugen’ (containing 1.775 items). We will focus on the narratives (‘beeldverhalen’) within these collections.

Current developments