Camera Self-Surveillance (2024)

8ch algorithmic video installation
PTZ IP cameras, crt tvs, custom code, object detection algorithm, Raspberry Pis, matrix switcher

Commissioned by WORM Rotterdam | Funded by Amarte Fonds

There are over 314.000 registered surveillance cameras in the Netherlands, as well as a large, indefinite amount of privately owned units - all collecting what can be described as "behavioural surplus data" on people, animals and objects. Most of these contemporary camera systems are running object-recognition algorithms, trained image-datasets, in order to send alerts when a person/intruder/object/etc is detected in frame. Strangely, the cameras themselves are almost never represented in public datasets.

The installation Camera Self Surveillance V1 is an experiment in collaboration with Christian Schwarz and Roel Weerdenburg to fill that dark-spot by creating an algorithm and dataset that tries to recognise surveillance cameras and uses them to observe one another. By turning the cameras back on themselves, the collective invites viewers to consider the implications of self-policing and how the normalisation of surveillance affects our behaviour. By focusing on how surveillance systems often operate unchecked, they challenge the assumption that these systems are in place solely for protection and control.

Image credits: Christian Schwarz