Collegium Helveticum
Fellow Project 2023–2024

Seeing (through) Images from Large Scale Vision Datasets

The rise of computer vision in society has gained prominence in recent years. Scholars often focus on the social impact of these algorithms, especially as they exacerbate injustices. However, these problems can only be understood—and avoided—if we look not only at the outputs of computer vision, but also at the inputs these machines were offered in the first place.

This emergence of computer vision coincides with the dissemination of so-called “Large Scale Vision Datasets,” short “LSVDs.” In these datasets, everything is intricate: millions of amateur images taken without consent from social media; low resolution files; all assembled through precarious labor carried out by thousands of micro-workers on remote platforms. Challenged by the opacity of such datasets, Bruno created a long-term methodological practice that aims at a deeper understanding of these images together with people directly or indirectly related to images, technology, and/or arts. The resulting material is invaluable for a more in-depth and less normative understanding of such images.

As an early-career fellow at the Collegium, Bruno’s intention is to make an experimental film that allows these training images to finally be seen. Academic outcomes (articles and attending conferences) will also form part of the results of this research.

Discover more