Dear Maria,
the relationship between the motion parameters and the resulting transformation matrices are different for FLIRT and eddy. I guess you are generating these transformation matrices as if these were parameters estimated by FLIRT?
Why do you want to resample the data using applyxfm4D instead of using the output images from eddy?
Jesper
On 2 Feb 2016, at 22:17, SUBSCRIBE FSL Anonymous <[log in to unmask]> wrote:
> Dear community,
>
> I've used the <basename>.eddy_parameters motion values in order to obtain the rotation and translation transformation 4x4 matrices for each of the DWI volumes. I was expecting that by applying these matrices with the applyxfm4D tool to obtain a motion-corrected set of images. However, when looking at the result, the data seems not to be corrected, as if, although the eddy tool has done a tremendous job motion-correcting the dataset, the transforms, that I have obtained for the motion and later applied on the motion-distorted dataset, do not give a motion-corrected result.
>
> Has anyone used the eddy_parameters for pure motion correction check? Is there any tool that correctly converts these parameters into the 4x4 transform matrix that is expected by the applyxfm4D tool?
>
> Thank you.
>
> Best,
> Maria
|