I actually wonder if we could exploit this ourselves somehow. Problem with this is it would likely require 3 different sensors each with a known distance from each other, scanning the same object. I don't know about the practicality of that.
cost would inevitably go up, but I don't think very much. maybe interlace 3 seperate sensors, each with a different polarization, then simply do the math.
Would likely still be MUCH cheaper than current high rez offerings, and likely much quicker too since much of the filtering is done in the physical domain.
CPU: Core i7
Projector: 3M Pocket Projector and LG PA75U