Morning Ian,
No personal experience with this
particular problem (instead, experience with vascular corrosion casting), but if
I were called, as you have been, I would try the following.
If you are being supplied with a
piece of skin, I would use a non-transparent meth-acrylic polymer, stained
after replication with a fluorescent dye that would give a strong emission
in the red or green. I would give the replica some thickness, i.e.,
1-2mm., then dissolve the last of the skin away (a la vascular corrosion
methodology) with 1N KOH (followed by considerable washing with distilled water
and critical point drying). A piece of the replica is
then mounted/suspended in 1.49 R.I. immersion oil in a chamber with an
acrylic coverglass bottom. Finally, view with a confocal system using an
inverted scope with objectives that are corrected for acrylic/plastic, and you
should be able to generate fairly true 3D images of the replica surface. I
would work towards using the smallest pinhole. Since the plastic would be
opaque, only the surface will provide signal, thus, the image stack should be of
contour lines only. The reconstruction should be a view of the surface of
the replica, AND, from the other perspective, a direct representation of the
topography of the surface of the skin. Every image can be viewed as an
altitude with precise separation from its neighbors, and it would be possible,
after some programming, to analyze each image in the stack for parallel
line pairs with measured lengths and separations (e.g., each wrinkle (crevasse)
in the surface will have depth, width and length available for automatic image
processing/analysis). In sum, the confocal, as you know, will provide much
better resolution on the z-axis for a study of this kind.
If the skin arrives still
attached to the patient, there are likely to be many methods for casting of
latex face masks that you can adapt, if the viscosity is sufficiently low to
provide the resolution you need. These materials and
methods will have already passed the toxicity and use tests that are
so arduous for patient safety measures. If the casting material is
non-transparent (as most of these would be), and fluorescent, the confocal
reconstruction method may be the most data-rich way to go. But, even
if the casting material cannot be safely made fluorescent before it is applied,
you can certainly stain it after removal from the patient.
There are software packages,
such as "ImageJ" from NIH (http://rsb.info.nih.gov/ij/) which
are not only free and open-source, but constantly upgraded with new capabilities
(including writing macros in Java) as well as plugins for lots of
morphometric analyses and the handling of stacks from most confocal
distributions. ImageJ also accepts other plug-ins that are
Adobe-compliant.
I offer this recommendation only
because I can't imagine trying to generate 3D information from the monster math
required when SEM images are the starting points. On the other hand, if
all you are going to do is count crevasses, perhaps that would be even more
simple/direct.
Hope
this helps,
Regards,
Fred
Monson
Frederick C. Monson, PhD
Center for Advanced Scientific
Imaging
Mail to Geology
West Chester University of
Pennsylvania
Schmucker II Science Center, Room SS024
South Church Street
and Rosedale Avenue
West Chester, PA, 19383
Phone:
610-738-0437
eMail: fmonson@wcupa.edu
-----Original Message-----
From: Ian
Montgomery [mailto:ian.montgomery@bio.gla.ac.uk]
Sent: Wednesday,
December 04, 2002 8:23 AM
To:
histonet@pathology.swmed.edu
Subject: Skin.
Have
been asked to make a replica of human skin then from the mask perform 3D
analysis of the lines wrinkles etc. Anyone have experience of this type of
thing. I'll do a web search but first hand experience would be
welcome.