| Realigning command in IDL. [message #14342] |
Thu, 18 February 1999 00:00  |
mph522
Messages: 2
Registered: February 1999
|
Junior Member |
|
|
Dear all,
I was just wondering whether there was a simple command in IDL/wave to realign two images together. I have two 3D MR datasets from the same subject which were taken six months apart, from which I'm trying to detect changes in brain volume (looking for voxels whose intensities have changed by a certain amount).
However, even after realigning and normalising the images using SPM96, the images are still not perfectly matched and this is leading to errors in my results.
Thanks in advance,
Carol.
|
|
|
|
| Re: Realigning command in IDL. [message #14418 is a reply to message #14342] |
Mon, 22 February 1999 00:00  |
David Foster
Messages: 341
Registered: January 1996
|
Senior Member |
|
|
Carol Docherty wrote:
>
> Dear all,
>
> I was just wondering whether there was a simple command in IDL/wave to realign two images together. I have two 3D MR datasets from the same subject which were taken six months apart, from which I'm trying to detect changes in brain volume (looking for voxels whose intensities have changed by a certain amount).
>
> However, even after realigning and normalising the images using SPM96, the images are still not perfectly matched and this is leading to errors in my results.
Carol -
2D images can be manipulated using POLYWARP() and WARP_TRI(), but since
your data is 3D you may want to try either:
1) AIR (Automated Image Registration) by Dr. Roger Woods of
the Laboratory of Neuro Imaging at UCLA:
http://bishopw.loni.ucla.edu/AIR3/
Source code is freely available, and it's supposed to work
fairly well.
2) You might want to get AFNI (Analysis of Functional Neuro Images)
by registering with the author, Dr. Robert Cox, at:
http://varda.biophysics.mcw.edu/~cox/index.html
This package is really meant for functional MRI, but it has a
3D registration program that is supposed to be much faster than
AIR and produce nearly identical results. It's free, but you must
register before downloading the software.
If neither of these solutions work for you or don't provide needed
functionality then email me and I'll explain what we've done here.
We're basically using a method similar to that used in AIR (known as
the Woods algorithm, using SVD to compute translations, rotations
and scaling required to align two series of tie-points).
Feel free to call me if you have any questions. Good luck! You might
consider segmenting your volumes into gray/white/csf and looking at
the totals of those compartments; this method would be less susceptible
to misregistration errors than trying to do a voxel-by-voxel comparison.
Dave
--
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~
David S. Foster Univ. of California, San Diego
Programmer/Analyst Brain Image Analysis Laboratory
foster@bial1.ucsd.edu Department of Psychiatry
(619) 622-5892 8950 Via La Jolla Drive, Suite 2240
La Jolla, CA 92037
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
