Re: Recognizing Aunt Tilly

[P3D Peter Davis <pfd@xxxxxxxxxxx>]

At 10:44 AM 10/31/96 -0600, you wrote:
>>The computer DOESN'T recognize Aunt Tilly, or Rover either. 
>
>PBS last week screened a documentary on the MIT Media Lab, where they're
>doing *exactly* that -- recognizing Aunt Tilly, Rover, and more. I'm a
>staunch believer in 2D/3D conversions. If I can *imagine* depth in a flat
>picture, then a human/computer collaboration could do the same. I believe
>Lincoln Kamm and others on this list have been fooling around with this,
>with some degree of success. 

The Media Lab did do some work with constructing scenes from views, but
these were not ordinary photographs.  They were video shot with a "range"
camera, which records distance to subject at each pixel of the image.

It is possible to construct 3D images from 2D using a computer, but it
requires a great deal of human intervention.  Essentially, it's like making
cardboard cutouts, but using Photoshop or some other software to manipulate
the images.  This is probably how the "Jurassic Park" View-Master slides
were made.

-pd

--------
                                Peter Davis
                        http://www.ziplink.net/~pfd/

                   "Nondescript -- the one word oxymoron."


------------------------------