Hello all,
I am attempting to use 3dAllineate to register a low-res (functional) spinal cord image to a high-res (anatomical) spinal cord image. The spinal cord structures are very small (1-2 mm) and thus the registration needs to be as accurate as humanly possible.
I commonly use 3dAllineate to register my brain images and it has worked extremely well, but I have experienced inordinate difficulty in getting it to work on spinal cord images. I have tried many different types of masks (binary, Gaussian of different sizes) with many cost functions (and other parameters) but nothing seems to work well (or robustly).
Has anybody successfully used 3dAllineate to register high-res (sub-mm) images? The data and code that I have been working with are available here (for the next 7 days) in case anybody is up for a challenge: [
vshare.vanderbilt.edu]
Any help or insight would be very much appreciated !
-Rob