"...reproduce "a replica accurate to within microns" of an original object snapped on a mobile device."
So, map your environment for a VR experience, but base it off an existing picture? Take a picture of an elephant, make a elephant with a 3D printer? Hopefully I'm guessing correctly and eventually it eliminates the temporary picture. Of course I'd trade the micron accuracy for a larger resolution, if possible. Not much is mentioned in the methods of down sampling or scaling with it comes to 3D image devices like these, but I hope there's options. Otherwise, it will sort of be like going everywhere with a 50mm prime lens, which isn't great, especially for 3D imagining (I won't be cuddling a Lion at 5 feet "for best results").