Holograms from CG 3D pictures. March 25 2004 at 6:11 AM
(no login)
You know, the ones you need to uncross your eyes and focus at a distant object. How can the information encoded in these pictures be extracted in holographic format such that a person does not have to learn the technique of uncrossing their eyes? It would be nice to be able to generate the 3D work on the computer, even a portrait as this site claims, and make a hologram of it. This will take some thinking!
This is an example of a 3D clue known as "disaprity. You may have seen that Joy is doing a talk on 3D vision at the PCG meeting in which she'll cover this. The first time I came across this was in Vegas about '90 or so. There was an ad about a "holographic display" in a Vegas mall. It turned out to be an ad for a mall stand in the middle of the floor selling these "Magic Eye" pictures. I thought then of actually making a hologram of them but discounted it. The problem with these Magic Eye pics is that they have no contrast, so that even if you see the 3D, it's difficult to tell what it is.
However, it's not difficult to make a HOE with which to see these.
I guess what I am grasping for is a method of combining all, many or some of the photos used in a stereogram into a single flat picture with wich you can make a hologram of and get a 3 dimensional image. Far fetched I know but it seems the algorithm exists for at least combining two.
Dinesh, is it possible to work out the math for combining, let's say, 5 photos all within the same master photo of the same size? Does this decrease resolution by 1/5? Then an HOE would break the master photo into its parts.
To encode a stereogram into a hologram is not too complicated. You simply need to make a hologram where each eye sees a different image. The problem is it only works from one exact view point.
