It looks as if DarbeeVision uses some form of advanced image stacking
to do their defocus and subtract to get the depth information. I
looked up their US patent # 7043074 and they do have very interesting
IP that applies to using image pairs, multiplets. One of their
accepted claims even covers aligned image pairs (wouldn't this be the
unsharp mask?). I have to go deeper into the image processing
functions, though the perceptor does a really good job of finding the
image detail. This entire package looks very good! I am very curious
how they are able to quickly synthesize a stereo pair, this is usually
very labor intensive.http://dvnphoto.com/downloads