the project is an exploration of algorithmically translating 2D visual data into 3D geometries at no control or resolution loss. various custom pixel projection workflows were tested for the final output, a spatial artefact that leverages deeplearning to embody collective memories of a space in the city into a physical object.

the question of why we collect, record, and share our quotidian experiences has always been entangled with the formal and aesthetic concerns about how to represent reality, totality, and the depth of human imagination. deep learning neuralnets were trained on 5000 geolocated images scrapped from the web. the network was used to synthesize new memories of viennas augarten and explore ways these "collective" memories could become 3D artefacts.

the morphing structure is a physical manifestation of citizens’ collective memory of the park. the columns below are early exploration into the potential of visual machine learning as a tool for space generation

in its final state, the design methodology relies on an algorithm which combines functional as well as cultural demands on the structure. the process-driven design constantly oscillates between physical and digital representations as well as rational and instictual methods of operation. the neuralnets visual information becomes graspable through translation into a walkable structure at this stage

the original visual input generating the complex geometric structure becomes revealed to the viewer through an object_tracking AR experience. cameras are used to access an overlay of the original synthetic hallucinations. the functionality of the interface is tested on a 1:10 prototype of the structure.