You must be logged in to post Login Register


Lost Your Password?
Search Forums:


 






Wildcard Usage:
*    matches any number of characters
%    matches exactly one character

Forum Diffusion Toolkit Usage Newbie question

Newbie question
Page:   1  
UserPost

10:32 am
July 27, 2010


Ish

New Member

posts 1
Print this Post
1

First I commend you for your software, I have been looking for a tractography tool and I find it very user friendly. I have couple of question thought.

1. Is this a probabilistic or deterministic algorithm?

2. I tried changing the parameter" Orientation patch, Mask threshold,Angle thershold" put my tracks still look very wrong. 

3. I usually get my gradient table from converting my dicom to niftii using dcm2nii. Do I have to modify this file for DTK?

Thanks for the response 

Ish

1:21 pm
July 28, 2010


Ruopeng

Admin

posts 353
Print this Post
2

Hi Ish,

To answer your questions one by one:

1. The alogrithm used in Diffusion Toolkits are all deterministic.

2. Can you post a screen shot and/or log text from DTK?

3. The bvec table pulled out of dicom header are localized vectors, while DTK takes the vectors in scanner coordinates. I just played a little bit with dcm2nii. If I use the bvec table it created, I had to apply 'invert z' to make tracks look right. Not sure why. I might investigate it further when I have time later on. Also you should uncheck 'oblique correction' even if your data is Siemens because the bvecs are the corrected ones.

Best,

Ruopeng

 

7:43 am
October 1, 2010


monicagiraldochica

Member

posts 8
Print this Post
3

Hi, I've done the same thing that Ish, I got my gradient table using dcm2nii, and after I read this post I applied invert z and unchecked 'oblique correction'. The tracks looked almost the same as before and I still cant make it run with the automatic image correction selected. Why can this be?

 

thanks,

Mónica

4:39 pm
October 1, 2010


Ruopeng

Admin

posts 353
Print this Post
4

I don't know what dcm2nii did but it appears to have lost the image orientation information during conversion. Have you tried to load dicom images directly to DTK?

Ruopeng

4:37 am
October 4, 2010


monicagiraldochica

Member

posts 8
Print this Post
5

Hi, well I tried to use directly de dicom images but they are in a folder that has more subforders with a lot of images, each one is one slice.

The thing is that another image that was corrected before using the DTK with the FSL does work perfectly.

Mónica

11:53 am
October 4, 2010


Ruopeng

Admin

posts 353
Print this Post
6

That's probably because FSL uses nifti format by default and keeps all the important header information. You should always use nifti format thourgh out your data processing routines if possible.

Ruopeng
Page:   1