The Wyckoff Principle
- The Wyckoff principle is based on the fact that differently exposed images
    of the same scene or object capture different information due to the
    differences in exposure.
    Consider, for example, these four images of the same scene: an open
    doorway leading off into a long dark corridor
    (captured with
    a 
    Personal Imaging camera, but the Wyckoff principle applies to
    equally well to ordinary hand-held 35mm cameras, HandyCams, etc.).
    
    
The above image shows the open doorway, upon which there is direct
     sunlight shining.  Since I am looking at the white sign on the door,
     my camera has adjusted itself to expose this nicely.  The
     door and wall are visible, but the inside of the corridor to the
     right is dark.
    
     
As I look to the right, the camera's automatic exposure
     brightens up the image so that now I can no longer read the sign.
     However, it's still not
     bright enough for me to see down into the dark hallway.
    
     
Now as I turn my head further to the right, I can begin to see
     just a little of the hallway.
    
     
As I look straight down the hallway, my camera's gain increases
     to properly expose what's in the center of it's field of view.
     I can see a person standing in the hallway (Charles Wyckoff,
     inventor of XR film).  However, the dark brown
     door to the left
     apears as completely white --- so white in fact, that we can no
     longer tell that it is a door.
    
    
    The entire dataset from this looking around
    is available to other researchers
    for illustrating/testing the Wyckoff principle
 - 
    Another similar dataset is also available.
     Here is a
    
    lookpainting
    generated, using the Wyckoff principle,
    from that dataset.
 
    
Notice how the sign on the door is very clear, and, at the same time,
    we can see down into the dark corridor.
 - 
    Joint parameter estimation in both domain and range of
    functions in same orbit of projective-Wyckoff group
    This paper pertains to creating high resolution radiance maps,
    by combining differently exposed pictures of the same subject matter
    in order to extend dynamic range.  (Based on an earlier
    1993 paper
    published in IS&T.)
 - A wearker result, for when the pictures differ ONLY IN EXPOSURE
    (e.g. when all the pictures are taken from the same camera position
    and orientation, as for example when the camera is mounted on a tripod
    and a number of pictures are taken with different shutter speeds)
    is described in the paper
    Being Undigital...
 - 
    A simple example of how the Wyckoff priniple can be applied to just
    two images (that's the minimum number required).
 - Here's another example of a Wyckoff set, consisting of
    some pictures taken at the cyberman
    opening at the Royal theatre.
    You can view or download
    the entire dataset from here.
 
--Steve Mann
I am a faculty member at University of Toronto;
you may want to visit my official faculty WWW page
at 
http://www.eecg.toronto.edu/~mann
Contact info: Professor Steve Mann, University of Toronto, Department of 
              Electrical Engineering,
              10 King's College Road, Toronto, Ontario, M5S 3G4
              
              mann@eecg.toronto.edu.
© We're supposed to have a
Copyright notice
in our Web pages.
That circle-C is ASCII 251 in case you were wondering.
My Snail mail (Canada Post)
is also on that same page.
 
Blue ribbon designates
 Free speech
 (ribbon used instead of
 setting background color to black
 because this would be a Notscape specific feature).
This page written for any browser, not
specific browsers.
It was tested with Mosaic (e.g. lean and mean, and COMPLIANT
to the HTML standards).