Computer Vision Art Gallery

CVPR 2021: Virtual due to COVID-19

Karen ann Donnachie & Andy Simionato

Regions of Interest, (2020)

Donnachie and Simionato use Machine Learning and Computer Vision to explore the early 20th century art techniques of collage, frottage, and decalcomania to generate a series of uncanny landscapes called Regions of Interest (2020). Their autonomous-art-systems create imagery based on computer vision, feature detection, and image analysis, mediated through a “calculated chance.” Specifically for this series, a custom-made algorithm selects and combines source images taken from digital archives such as open source libraries, web-cams and video-feeds, using a ‘salience matrix’ in order to create unique photographic collages. The source imagery used to create each collage is “cut” and combined, according to algorithmic gestures, originally human-taught and increasingly conditioned through the procedural processing of the archive via Computer Vision and Machine Learning. The resulting otherworldly scenes cannot be visited, they occupy a mindscape somewhere between the familiar and alien. For now, they are simply classified by the artists as “regions of interest” to be archived in a rapidly growing atlas of nonhuman destinations. The extension of this project entitled The Age of Reason attempts to automatically impaginate and publish the collage outcomes in souvenir picture books with an additional paratextual element of nonhuman typography, also subjected to the same algorithms of collage, frontage, découpage and decalcomania.

Back To The Gallery