[IGSTK-Developers] Re: segmentation techniques for the liver

[Stephen R. Aylward]

I've uploaded liver CT scans from 3 patients to
http://caddlab.rad.unc.edu/Public

The files won't appear in the listing generated at that page.  You must 
type the filenames into the webpage path and save them to your disk. 
The files are
Liver072.tar.gz
Liver073.tar.gz
Liver080.tar.gz

Each file is about 50 megs.

Each collection has several 2D scout images and two 3D volumes.  The 
volumes were acquired at two different time points as the contrast agent 
passed thru the liver - in the first the contrast was primarily in the 
portal venous system - in the second the contrast also includes the 
hepatic venous system and has begun to wash out from the portal venous 
system.  For liver segmentation, use the second 3D volume - the liver 
parenchyma will have taken up the contrast and it will thereby have 
better contrast vs adjacent structures.

These data are old, but good quality (1mm x 1mm x 3mm - clinical 
standard at UNC).  These patients were recruited using old HIPAA 
requirements and IRB guidelines, and their data was anonymized without a 
way of tracing back to the patients.   I discussed this with UNC's IRB 
office, and I can release these data; however, please do not distribute 
these data to other groups - I am still trying to keep track of their 
use and distribution just in case IRB changes their minds.

I will remove these files within a week.  Get them now if you want them.

Stephen

Kevin Cleary wrote:
> Stephen:
> 
> This is great stuff, and I will pass it along to my colleagues working on
> this part of the grant (Ziv Yaniv in our group, my clinical colleague Fil
> Banovac, and Steve Haker and Noby Hata at BWH) and get their input.
> 
> I would love to get the sample CT data sets as well just for the guys in the
> lab to start trying a few things
> 
> Kevin
-- 
=============================================================
Stephen R. Aylward, Ph.D.
Chief Medical Scientist
Kitware, Inc.
http://www.kitware.com