Watching Historical Films Through AI: Reflections on Image Retrieval from Heritage Collections

Abstract

Computer tools allow us to watch films differently and offer film scholars the opportunity to ask diverse research questions and imagine innovative methods. This article focuses, more precisely, on how film historians could benefit from Artificial Intelligence (AI), a tool which has not been used much in the field. The objective is to share a first round of experiments using Snoop, an AI developed by a collaboration between the French National Audiovisual Institute (INA) and the National Institute for Research in Digital Science and Technology (INRIA), on a corpus of historical films that, because they have been aired on television, are available for research from the INA collections. Following an exploratory approach, Snoop allowed us to retrieve shared motifs in a corpus of historical films. In addition to discussing challenges and findings, this article aims to evaluate the overall experience and open new lines of inquiry. From a historiographical point of view, we aim to assess whether, and if so, how, AI systems are useful tools for the implementation of different methods that both answer new research questions and revisit old ones.