Segmentation.Network – Projection

found dataset, custom software, raspberry pi, projector, 2018

Exhibited at Kunstverein Hildesheim as part of “Performing the System”, curated by Luzi Gross (in the front: an early version of “Hopes and Deliveries”) Photo: Kunstverein Hildesheim

Exhibited at Drugo More as part of “Who does what?”, curated by Silvio Lorusso. Photo: Tanja Kanazir / Drugo more

Exhibited at Drugo More as part of “Who does what?”, curated by Silvio Lorusso. Photo: Tanja Kanazir / Drugo more

Exhibited at Drugo More as part of “Who does what?”, curated by Silvio Lorusso. (bottom front Constant Dullaart, bottom back Anxious to Make, on the left Ottonie von Roeder)

Segmentation.Network plays back over 600.000 segmentations manually created by crowd workers for Microsoft’s COCO image recognition dataset. This dataset is based on photos from Flickr and is used in machine learning for training and testing.

The piece makes visible the hidden manual labor that goes into building neural networks and artificial intelligence. Furthermore, Segmentation.Network addresses machine vision as an act of conscious selection: what can and should be seen by machines and what will remain unrecognised or deemed irrelevant is separated by distinct lines.

Thus, Segmentation.Network considers neural networks and artificial intelligence in general as a collective and rather introspective endeavor and achievement.