Files in these directories are were generated from the validation
study of watershed segmentation done at the University of Utah.
Some of the hand segmentations were done by the Surgical Planning Lab 
at Harvard Brigham and Women's Hospital.  See the documentation for 
Validation/WatershedSegmentation in the InsightDocuments repository available
from www.itk.org.

All files are unsigned char, 8 bit data.  Pixels with 0 values are background
pixels.  Pixels with nonzero values are foreground pixels.  Values of
foreground pixels may vary from file to file, but are generally 1.  Endianness
of the data is irrelevant.

The EYE, RECTUS, and LENS data sets are registered to 

VisibleWomanHead/256Nraw

Available from the Insight data repository.


The OPTIC data sets are registered to 

VisibleWomanHead/256OPTICraw

Available from the Insight data repository.

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

EYE.  These data sets are denoted by the letter "e", as in 10e.raw.cropped,
etc. 
SIZE: 103 x 103 x 80
REGISTRATION OFFSETS (x, y, z): (+1, +132, +319)

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

RECTUS. These data sets are denoted by the letter "r", as in 8r.raw.cropped,
etc.
SIZE: 97 x 114 x 35
REGISTRATION OFFSET (x, y, z): (+7, +110, +342)

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

LENS. These data sets are denoted by the letter "l', as in 1l.raw.cropped, etc.
SIZE: 32 x 18 x 49
REGISTRATION OFFSET (x, y, z): (+34, +197, +335)

----------------------------------------------------------------------------
OPTIC. These data sets are denoted by the letter "o", as in 5o.raw.cropped,
etc.
SIZE: 255 x 255 x 53
REGISTRATION OFFSET (x, y, z): (+2, +2, +1)

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