Guided by the motto "We make animation visible!", the digital humanities project AniVision uses machine learning and computer vision to explore a hitherto marginalized part of the stylistic history of animation. The project focuses on ephemeral films such as educational films, commercials, or public service announcements. Animation is used in a variety of ways in these films, e.g., to advertise a product in a memorable way or to effectively depict a process inside the human body. AniVision investigates these different uses and stylistic patterns of animation by analyzing a large corpus of ephemeral films produced during the Cold War period in Austria, West Germany, and East Germany.

The project follows an interactive, computer-assisted approach that closely integrates animation research and computer science. The computer scientists develop automated corpus exploration techniques that the animation scholars then employ for the qualitative and quantitative analysis of the material. The computer scientists use methods from computer vision, content-based image and video retrieval, and interactive machine learning to extract syntactic and semantic attributes from animated footage. The animation scholars combine neoformalist film analysis with quantitative methods and other approaches from animation studies and art history.

AniVision is a transnational collaboration between Dr. Erwin Feyersinger (University of Tübingen, Germany), Dr. Franziska Bruckner (University of Applied Sciences St. Pölten, Austria), and Dr. Matthias Zeppelzauer (University of Applied Sciences St. Pölten, Austria). The project’s three-year running time (2023–2025) is funded by the FWF and DFG as part of the Weave Lead Agency Procedure. Claudius Stemmler works in the project as an academic researcher at the University of Tübingen. You can find further information on the project webpage at the University of Applied Sciences St. Pölten.