T3D Re: TECH-3D digest 337

[Rolf Henkel <henkel@xxxxxxxxxxxxxxxxxxxx>]

Well, many thanks to John B.:

> (Theory Of Everything Stereovisual) also known as
> "Stereovision by Coherence-Detection".

:-)  - but I guess it's fair to say that we are far away from a real
understanding of how stereovision works in our brains.  Some additional
comments on numbers:

Hyperacuity (or Vernier acuity, as it is sometimes called) shows up in a
lot of visual tasks. You get discrimination performance between 5 to 20
secs, depending on what type of task you are measuring. Interestingly
enough, stereovision is intrinsically working with hyperacuity - a fact
which brought me on the track with my own theory in the first place.

I cite from a great (but sadly quite expensive) book, "Binocular Vision
and Stereopis", I.P Howard & B. J. Rogers, Oxford University Press 1995,
p. 155: "There is general agreement that, under the best conditions,
stereoacuity is in the range 2 to 6 arcsec. This is similar to the limits
of vernier acuity."

This number has to be compared to 1 min, which is more or less the visual
acuity the receptor-spacing in the fovea would suggest. So stereovision
can be 20 times more precise! Btw: the above numbers would correspond to
a power of depth-resolution of about 3-7cm at a distance of 10m, and
6-15mm at a viewing distance of 5m (according to another source). This
precison of the human visual system is always amazing for me!

Rolf H.


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