Hi Ed,
This sounds like a good solution to increasing your SNR
for the eddy-correct. You are definitely right that you
should then apply the matrices to the unfiltered data.
It is pretty easy to split up the ecclog file (which contains
the matrices) or modify the script to save them directly.
Either way, if you have any trouble with this, or applying
the matrices, just let us know.
All the best,
Mark
On 8 Mar 2012, at 16:40, Ed wrote:
> Hi Mark,
>
> I've just found a solution: I filtered the DTI data with
> a 2D median filter with a 3 x 3 kernel in advance. The
> result is a near-perfect eddy-current correction. The
> only thing to do is to save the output matrices and
> apply these to the unfiltered data. But that's
> straightforward.
>
> Regards,
> Ed
>
>
|