BDP Windows vs. Linux

Viewing 6 reply threads
  • Author
    Posts
    • #5369
      waycon
      Participant

      Hello, I note that BDP works nicely on Windows 11 with the sample files : 2523412 for nii , bval and bvec in Windows using :

      “C:\Program Files\BrainSuite23a\bdp\bdp.exe” 2523412.bfc.nii.gz –tensor –FRT –FRACT –nii 2523412.dwi.nii.gz -g 2523412.dwi.bvec -b 2523412.dwi.bval

      : but when I try the equivalent in Linux Mint (Ubuntu), using :

      : /home/wayne/BrainSuite23a/bdp/bdp.sh /home/wayne/Desktop/mrtest2/2523412.dwi.bfc.nii.gz –tensor –FRT –FRACT –nii /home/wayne/Desktop/mrtest2/2523412.dwi.nii -g /home/wayne/Desktop/mrtest2/2523412.dwi.bvec -b /home/wayne/Desktop/mrtest2/2523412.dwi.bval

      : (I’ve altered the directories in BDP.sh to reflect my file structure), I get this error message:


      Setting/generating masks…
      Matching centroids (approx. align)…Warning: Matrix is singular, close to singular or badly scaled. Results may be inaccurate. RCOND = NaN.
      > In register_files_affine (line 329)
      In EPI_correct_files_registration_INVERSION (line 134)
      In coregister_diffusion_mprage_pipeline (line 246)
      In BrainSuite_Diffusion_pipeline (line 57)

      Warning: Matrix is singular, close to singular or badly scaled. Results may be inaccurate. RCOND = NaN.
      > In register_files_affine (line 330)
      In EPI_correct_files_registration_INVERSION (line 134)
      In coregister_diffusion_mprage_pipeline (line 246)
      In BrainSuite_Diffusion_pipeline (line 57)

      Matching resolution of the images…Warning: Matrix is singular, close to singular or badly scaled. Results may be inaccurate. RCOND = NaN.
      > In reslice_vol_same_grid>get_voxel_grid (line 160)
      In reslice_vol_same_grid (line 59)”….

      : I generated the bfc by progressing through the Anatomical Pipeline. Does anyone have any ideas about this inconsistency? (P. S.: I’m starting this string because I couldn’t seem write to the BDP questions I found elsewhere on the forum. I hope you will accept my apologies in advance.) wc

      • This topic was modified 3 weeks, 2 days ago by waycon.
    • #5375
      David Shattuck
      Keymaster

      hi –

      It looks like you are running this on our sample data. We use those data routinely to test for cross-platform consistency, so we haven’t encountered this issue.

      I do see some differences in the inputs between the two calls you make.

      Your file on the linux line is named “/home/wayne/Desktop/mrtest2/2523412.dwi.bfc.nii.gz “, but the call in Windows uses 2523412.bfc.nii.gz. What is /home/wayne/Desktop/mrtest2/2523412.dwi.bfc.nii.gz? Is it the same as 2523412.bfc.nii.gz or did you apply bfc to the dwi data?

      The other difference I see is 2523412.dwi.nii.gz vs 2523412.dwi.nii, but that shouldn’t affect anything unless the files don’t match.

      thanks,
      David

    • #5380
      ChrisS
      Participant

      Thank you for responding, David,
      I was delayed by a computer crunch and by business. To answer what I believe is the main question, the bfc file I used was created by myself from the processing by the BrainSuite GUI extraction process, rather than the one that came with the other files. Is that an error?
      I thank you for your time,
      Wayne
      P.S.: I’ve log in on another account shared by myself and another. wc

      • This reply was modified 1 week, 5 days ago by ChrisS.
    • #5385
      David Shattuck
      Keymaster

      hi Wayne –

      I think that could be it. Is 2523412.dwi.bfc.nii.gz a T1 and not a DWI image? The name suggests it was created from 2523412.dwi.nii.gz. I would check that first.

      I would also just copy the files that worked as inputs on Windows over to your Linux machine and test those.

      Thanks,
      David

    • #5387
      waycon
      Participant

      Hi David,
      I was finally able to login as myself again. WordPress’ security is tight.
      I’ve done a reinstall of BraineSuite23a and MATLAB19a on both of my computers and both crash when I get to the Pial Surface step. Now, the documentation seems to be directing me to different versions of MATLAB Runtime. Would I get as far as the Pial Surface step, if I had the wrong version? Do I need more than one version? R2019a? R2023a? I’ve installed, then uninstalled each. I’m guessing that, in part, I haven’t located the most recent documentation. Do you have a moment to give me any guidance? I would understand, if you don’t. I hope you don’t mind me asking.
      In any case, I thank you for your time,
      Sincerely,
      Wayne

    • #5388
      David Shattuck
      Keymaster

      hi Wayne –

      We list the requirements on the download page.

      BrainSuite23a requires MATLAB 2023a or MATLAB Runtime 2023a. You don’t need other versions of the runtime if you are using BrainSuite23a.

      Versions 19b through 21a required MATLAB (Runtime) 2019b
      Versions 16a through 19a required MATLAB (Runtime) 2015a.

      We’ll be moving to MATLAB 2024b for our next release (hopefully this month) because it provides improved support for Apple silicon and a number of other features we plan to use in the future.

      The cortical extraction sequence doesn’t require the MATLAB libraries. If you are running the full anatomical pipeline, then the MATLAB dependencies begin with cortical thickness estimation. If the program is having trouble prior to this (e.g., if it isn’t generating the pial surface), then the problem isn’t MATLAB.

      Can you provide more detail on how you are running the programs? Are you using the GUI or command line?

      thanks,
      David

    • #5390
      waycon
      Participant

      Hi David,
      I’m excited to hear there will be a new version coming out soon. As for my troubles, they are the same when I run the GUI either with the Anatomical Pipeline or stepping through. I start with a DWI with bval and bvec in a dedicated folder. When the Pial Surface step begins, a small black box pops up and the entire BrainSuite program shuts down. With terminal, I have been able to run bdp successfully, if I somehow get the correct bfc, but the bdp program often rejects the bfc file created in the earlier stages of the Pipeline. The error message includes the phrase “Matrix is singular.” Also, when running bdp, I don’t have a label file for the process, I presume because that file is created in the pial step.
      Is this what you need? And, again, let me thank you for your time.
      Sincerely,
      Wayne

Viewing 6 reply threads
  • You must be logged in to reply to this topic.