How do I extract the diffusion tensor from DICOM files for use in a math modelling project, either in MRTrix or MATLAB? We want the eigenvectors for each voxel from the diffusion tensor for our modelling equations. Does anyone have experience in creating a matrix of eigenvectors from this data or any way of efficiently extracting the diffusion tensor? Thanks.
William,
You should be able to convert the images to nifti files using dcm2nii. From there I would recommend taking a look at the DTI tools that are part of the FSL data analysis suite. After fitting the DTI data you will end with 3D files containing the information you need. You can then read these images into Matlab if you need to. FSL also has nice registration tools and other bells and whistles you may want to use.
I have been converting them into nifti files to view in MRTrix, my colleague says you lose some data, for example with our HARDI images, which use a 64 direction approach to get the tensor. Have you experienced this? For starters I will try these techniques as a simple diffusion tensor is all that is really needed to start the modelling so if we lose some image data at this point it is not vital. A MATLAB output would be perfect though, I will give these a try, thanks.
Brilliant, thanks Sabine, my colleague has apparently managed to extract the diffusion tensor in MRTrix, but this is not directly exportable to MATLAB without some extra processing as far as I am aware. I downloaded 3D Slicer the other day, but have been flat out cramming for my medical exams. I will try this myself and see if I can get it working as that would make things _much easier_
- Badges
- Science topic