It would be useful to reduce bandwith demand of HD cameras. You'd need 3 cameras, reduce the image depth to 4 bit which would lower the background to basically clipart quality, create geometric objects from it and reduce bitmap colors down to 16 colors.
Create a fractal of the basic background geometry and imbed that formula into the recorded video stream which is processed only on playback. Motion areas remain in 1080p but the rest of the scene is like a cartoon.
Photo snaps, once a second, if detail is needed, in a separate file. 3D scene exported to .dxf or sketchup.
If you're not doing it I will