[Camino-users] problems creating .Bfloat file

[Philip A Cook]

Assuming the first voxel (usually nothing interesting) shouldn't be zeros, I would suspect something odd with the header and / or the input data. I'm assuming that you've looked at the DWI images and they are not in fact zero. Can you do 

niftiheader -readheader dwi.nii.gz 

and send me the output?


On Mar 22, 2013, at 4:06 PM, Christina Hugenschmidt wrote:

> Thanks Philip - I ran the command you put below (image2voxel -4dimage dwi.nii.gz -outputfile data.Bfloat) and it successfully created the file. However, the Bfloat file is populated with 0s (if I run the command cat file.Bfloat | float2txt | head -50. When I looked at the first Bfloat file I made more laboriously, the same was true (the Bfloat file is 0s). Do you have any idea why this might be? 
> 
> Christina E. Hugenschmidt, PhD
> Instructor
> 
> Section on Gerontology & Geriatric Medicine
> Medical Center Boulevard \ Winston-Salem, NC 27157
> P 336.713.4190 \ f 336.713.8588
> chugensc at wakehealth.edu
> 
> 
> 
> -----Original Message-----
> From: Philip A Cook [mailto:cookpa at mail.med.upenn.edu] 
> Sent: Friday, March 22, 2013 3:18 PM
> To: Christina Hugenschmidt
> Cc: camino-users at www.nitrc.org
> Subject: Re: [Camino-users] problems creating .Bfloat file
> 
> Hi Christina,
> 
> If you are using dcm2nii it should be possible to convert your data and scheme file in one step apiece.
> 
> For the data:
> 
>  image2voxel -4dimage dwi.nii.gz -outputfile data.Bfloat
> 
> For the scheme file:
> 
>  fsl2scheme -bvecfile image.bvec -bvalfile image.bval -bscale 1 > scheme
> 
> Here I've used -bscale 1 so that b-values will stay in units of s / mm^2. This is OK for most DTI uses but it will be an issue if you run dtlutgen or datasynth - in which case you will need extra options to let those programs know that you are using s / mm^2
> 
> Then you should be able to do
> 
>  modelfit -inputfile data.Bfloat -schemefile scheme -model ldt -bgmask mask.nii -outputfile dt.Bdouble
> 
> 
> If you have a recent version of Camino you can also import NIfTI data directly
> 
>  modelfit -inputfile dwi.nii.gz -schemefile scheme -model ldt -bgmask mask.nii -outputfile dt.Bdouble
> 
> This will use a bit more RAM though
> 
> 
> On Mar 22, 2013, at 2:29 PM, Christina Hugenschmidt wrote:
> 
>> Hi,
>> 
>> 
>> 
>> I am new to camino and quite a novice w/r/t programming. I was able to successfully run the tutorial on the human tutorial data. With those disclosures, here is my question. The first time I tried to create an FA image on my own data, it was empty. After emailing with Uran and reading the data import page, I decided the problem is likely that my data are in the wrong format. I have data acquired on a 3T Siemens scanner (30 directions + 1 B0) that are 16 bit signed integer and little endian.
>> 
>> 
>> 
>> I followed the instructions on the data import page (specified the size of the data set, stripped the nifti header to make a raw data file, and used the scanner2voxel and shredder command at the bottom of the page). I had already made a scheme file from the bval and bvec files created by dcm2nii. The scheme file seems to be correct when I look at it, so I tried again to run the tutorial. When I ran the command modelfit -inputdatafile file.Bfloat -schemefile file.scheme - model ldt -bgmask file_mask.nii -outputfile dt.Bdouble, I got the warning "couldn't parse arg 0: '-inputdatafile'" then "couldn't parse arg 1: 'file.Bfloat'".
>> 
>> 
>> 
>> I'm not sure what I'm doing wrong, and I would appreciate any insights.
>> 
>> 
>> 
>> Thanks,
>> 
>> Christina
>> 
>> 
>> 
>> Christina E. Hugenschmidt, PhD
>> Instructor
>> <image001.gif>
>> Section on Gerontology & Geriatric Medicine Medical Center Boulevard \ 
>> Winston-Salem, NC 27157 P 336.713.4190 \ f 336.713.8588 
>> chugensc at wakehealth.edu
>> 
>> 
>> 
>> _______________________________________________
>> Camino-users mailing list
>> Camino-users at www.nitrc.org
>> http://www.nitrc.org/mailman/listinfo/camino-users
> 
>