image registration

[hermione]

Dear fellow researchers,

I am a novice in the field, so please forgive my ignorance (: 
I am working with fMRI data, and would like to "standardize" the voxels across participants before I perform my analysis.  I hear that registering the voxels to the standard Talairach image is one way of doing it.  Could anyone here suggest a free software that works on Windows and with matlab files that can do that?  I would be tremendously grateful for any other suggestions, advice, etc.

Thanks!

[Xenios Papademetris]

BioImage Suite can do these things .... you will need to save your data out of Matlab into Analyze format (by early next month we will also be supporting the NIFTI format).

A good primer for registration for the issues (I think) is a tutorial I gave at Yale a c ouple of years ago, see: http://noodle.med.yale.edu/~papad/various/papademetris_Image_Registration.ppt

Good luck.

Xenios

[hermione]

hi again,

this is probably a stupid question, but how do I convert the data from matlab to analyze format?  BioImage Suite is available for free, correct?

Thank you.

[Xenios Papademetris]

BioImage Suite is available for free!

Tools to read/write Analyze and NIFTI format (we will be supporting this in the next cut) can be found at:

http://www.mathworks.com/matlabcentral/fileexchange/loadFile.do?objectId=8797&objectType=FILE

Xenios

PS: A Google search using the terms "Matlab" & Analyze found this. Also see

[Xenios Papademetris]

The talks were moved to a different location during a webpage reorganization.

http://bioimagesuite.org/talks/Introducing_BioImage_Suite.ppt Introducing BioImage Suite.
http://bioimagesuite.org/talks/dti-talk.ppt DTI Talk.
http://bioimagesuite.org/talks/rajeevan_fmri_statistical_analysis.ppt FMRI Talk.

The fMRI statistical analysis tools are not yet publicly available, hopefully this is coming soon.

Xenios

[hermione]

Hi,

I was using intersubject registration in bioimage suite and the output files came out in MATR format.  Could you please explain this format?  I have never heard of such a format, and even bioimage suite didn't recognize the files when I tried to load them.  Is there a way to convert them into Analyze format?

Thanks.

[Xenios Papademetris]

The output format is simply the matrix mapping one image to the other. You can use this to reslice the images. See

http://bioimagesuite.org/public/CommandLine.html#reslice

Xenios

[hermione]

I have another quick question--what's the difference between intensity-based and point-based methods of registration?

[Xenios Papademetris]

Intensity based registration uses the image intensities directly -- this is the most common form and the most automated. Point-based (feature-based) methods first require extraction of key landmarks/surfaces etc. and then use these for registration. They (point-based) can be more accurate for specific regions but require additional effort to extract the surfaces. (One example where this works really well is in mouse-CT registration where the bone surface can be cleanly and easily extracted)

Xenios

PS: See http://noodle.med.yale.edu/~papad/various/papademetris_Image_Registration.ppt for a quick overview talk on Registration.