Ongoing and emerging research:
Preliminary experiments into training an AI machine with Google Earth image sets. This latent video morphs between different images created by the machine from the Google Earth image references.
'Monocular Depth Estimation' - AI model estimation of depth in 3D space captured in still images. Shown here
on image of a diver scanned from 1978 National Geographic 'Dragons on the deep' issue.
Further monocular depth estimation experiments
'In Image Painting', this model is used to 'repair' damage or replace missing areas in paintings or images.
Here they have been used to manipulate and attempt to repair shots taken from Google spheres that have ruptures or glitches.