Hi. I'm trying to process a diffusion data, from an rat brain acquired from a Bruker scanner. 
I have manually changed the file order, B0, B1000, and B1600 as mentioned in the FAQ and documentations. 
I split the files, removed 4 of the 5 B0 images, and re-ordered them in a way that matches the provided sample image. 
I have a 30 direction gradient vector, which makes the process run without problem. 

However, the processed maps, they have some problems. 
Can anybody help me with this issue, or does anyone similar experiences and have manage to get the things going right? 

Thanks for the time spent reading this question.

Dear Ho-Joon, 

I apologize for the delay in answering this question. How high is your background threshold T?

T = 50; % background threshold (default: 50); voxels with b = 0 intensity values above this threshold (after optional connected component analysis) are processed

Voxels with an intensity value above this threshold T will be removed and result in a value of zero.  Additionally, in stead of removing 4 of the 5 B0 images, it is more common to create an image that is an average of all 5 images. 

If the background threshold is not what is causing your problem, please respond with your dkeparameters file, this will help me in finding another possible cause for your problem. 

Best, 
Emilie

Hi,I am facing the same problems,do you now have some ideas about how to resolve it?I think there may be soming wrong with the grident table,but I'm not sure.I receive the same result as you using my own gridents table(get from dcm2niigui),while it's ok when I use the grident table of DKE,only the quality of the image is very bad.
if you have the right methods,please give me an response.I would be thankful for that.

Hi Shi Li, 

If you used different gradient directions during aqcuisition than the ones incorporated in DKE (those are the standard Siemens Tim Trio tables), it is important you use your own gradient tables. After converting your data with dcm2nii you have to transpose your matrix before importing it in DKE. Specifically, dcm2nii outputs a table that is 3xN and DKE expects a table that is Nx3. If your problem persists, please provide us with your DKEparameters file and an example output map. 

Best, 
Emilie