The repository contains the implementation of an enhanced version of the FSL eddy correct tool that is designed to reduce registration bias observed at b=3000. The method was described in:
Improved Accuracy of Motion and Affine Eddy Current Distortion Correction in High B-Value Diffusion Weighted Imaging Using Brain Mask Based Weighting Functions, Christopher Leslie Adamson, Richard Beare, Deanne Thompson, and Marc Seal, ISMRM 21st Annual Meeting and Exhibition, April 2013.
DWIEddyCorrectMI is a python script with the following dependencies:
- python 2.7 and the following packages:
- numpy
- scipy
- nibabel
- distutils
- FSL 5.x
- GNU Parallel (optional)
DWIEddyCorrectMI [-p] [-m <output mat file>] <input file> <output file> <reference index>
Performs alignment of each volume in <input file> to the image in the input file specified by <reference index>; the first volume is 0.
- -m : contains all of the transformation matrices concatenated
- -p use GNU parallel to perform the registrations in parallel, uses GNU parallel