Automated Processing of Auction Catalogs

This lecture presents the internship report of Raphaël Barman, a digital humanities master student at EPFL, who worked on automatically processing information from auction catalogs during a five-month internship in Paris. The project focused on creating a pipeline to digitize and extract structured data from auction catalogs, aiming to analyze the content digitally. Barman used image classification and segmentation techniques, including neural networks, to extract objects and artists from a subset of around 2000 catalogs. The internship provided valuable practical experience and the opportunity to collaborate across different domains, highlighting the application of digital humanities in real-world scenarios.