Possible Bodies

Panoramic Unknowns

Item number: 108
Item title: Panoramic Unknowns
Entry of the item into the inventory: 28 January 2019
Activities in which the item participated: Optimization and its miscontents
Inventor(s) for this item: Possible Bodies

Unknown Object Tracking [1] is being applied on an industrial scale in robotics, car traffic control and 'intelligent' surveillance systems; think Amazon warehouses and Google street view. This technique employs panoramic cameras, which provide 360-degree views that are processed by specialised software in order to locate volumes in space. But when surveying areas like factories, big halls, streets or even a single room, there are always many 'blind spots' and 'unknown objects'. Machine Learning is currently being employed to optimize such known unknowns.

Possible Bodies [2] is a collaborative research on the very concrete and at the same time complex and fictional entities that "bodies" are, asking what matter-cultural conditions of possibility render them volumetrically present. For the workshop Optimization and its miscontents [3] we invite participants to use their political/aesthetic/technical sensibilities towards the computation of volumetrics. We invite you to attend to Unknown Object Tracking [UOT] as a spacetime mattering that is not necessarily organized from the outside to the inside (as the optimised god-like sight) but from the inside-out, somehow both adjusting to and reversing a 3D paradigm of convergence into a "point of origin".

The seamlessness that a 360 capture demands is based on the analysis of tiny 'features' that are sophisticatedly attended to by Computer Vision agencies. Features are differences, anomalies, visual anchors, identifiers that are considered to be specific for a single capture.

UOT depends on processing past data and simulations that are kept, learned and trained before being projected into the future, while real time data from the camera is used to adapt to changes. Their predictive claim is based on past measurements that make optimization systems vulnerable to unexpected change and/or wild surprises, as the unknown can only be detected if it falls within the boundaries of the probable. Where is the possible in considering panoramic unknowns? What if we consider the unknown not only as risky but fundamental to an orchestation of spacetime that is not necessarily harmonic and predictable, but also dissonant and accidented?

[1] https://vision.unipv.it/CV/materiale2016-17/4th%20Choice/0293.pdf

[2] http://possiblebodies.constantvzw.org

[3] https://2019.transmediale.de/content/optimization-and-its-miscontents-counterpolitics-of-surveillance-capitalism

https://encrypted-tbn0.gstatic.com/images?q=tbn:ANd9GcQsZUtukHOTeGWoJEzbVqliVjf6h0MNwTwsXs7JFOwIYdqDHl5Cgw https://laspromesasdelosalgos.hotglue.me/features_pixels