From: brucec@phoebus.labs.tek.com (Bruce Cohen)
Subject: Re: head tracking
Date: 2 Dec 91 22:12:36 GMT
Organization: Computer Research Lab, Tektronix Inc.



In article <1991Dec2.152517.17607@milton.u.washington.edu> jim@baroque.
stanford.edu (James Helman) writes:

>    2) Head mounted cameras tracking a set of LEDs in the ceiling.
> 
> There's no comparison.  #2 give very big motions on the image plane.
> But in order to get any kind of angular resolution out of #1, you'd
> need to have cameras which could zoom and pan to track the head, or
> lots of fixed cameras with narrow fields of view.  Otherwise, the
> change in positions of the LEDs on the image would be too small for
> any kind of decent resolution.  Also, if you want wide area tracking,
> you'd need lots of cameras in the ceiling.  LEDs are cheaper.

OK, but couldn't you get even more change by mounting lots of LEDs in
the room, along with a few cameras, and then putting mirrors on the
user?  (ignore my previous posting about retro-reflective mirrors; they
won't help with the field of view problem).  If a mirror rotates by N
degrees, the beam of a given LED reflected in that mirror sweeps 2N
degrees, twice was much as a camera or LED would.

------------------------------------------------------------------------
Speaker-to-managers, aka
Bruce Cohen, Computer Research Lab        email: brucec@crl.labs.tek.com
Tektronix Laboratories, Tektronix, Inc.                phone: (503)627-5241
M/S 50-662, P.O. Box 500, Beaverton, OR  97077