- This topic has 14 replies, 2 voices, and was last updated 9 years, 2 months ago by
Anand Joshi.
-
AuthorPosts
-
-
August 16, 2016 at 1:16 pm #294
Hi Francois,
Thank you for reporting this. Did you get similar errors on other datasets (if you have processed other brains) with BrainSuite or only this one?
Anand -
August 16, 2016 at 1:20 pm #295
It’s the only one I tried…
-
August 16, 2016 at 3:20 pm #296
Hi Francois,
BrainSuite+SVReg is running fine on my computer so far. It has passed the stage where it showed error.
I am working with sample_auditory dataset. I downloaded it from brainstorm site, created a new subject in brainstorm and imported anatomy for sample_auditory there. Then I exported the mri as T1.nii
I ran brainsuite on this data with default parameters, except at skull extraction, I chose Edge constant=0.63 and Erosion Size=2.
Can you please make sure that this change was done when you ran the sequence? It could have caused bad extraction and then error later.The following was mentioned on the BrainSuite tutorial webpage on the BrainStorm site but now seems to be removed in the subsequent edits. I will reinsert it. Hope that it is fine.
For this tutorial data, the parameters for skull extraction need to be changed to get optimal results. Please set ‘Edge Constant=0.63’ and ‘Erosion Size=2’.
Thanks,
Anand
-
-
August 17, 2016 at 7:13 am #297
Hi Anand,
I got exactly the same error with the default values and with Edge Constant=0.63 and Erosion Size=2.
Except that in the second case (modified values), the pole of the frontal lobes is missing.
I removed this line from the tutorial because modifying these values was giving results that were worse than with the default values.The file that should be processed in this tutorial is the following (I fixed the download link):
ftp://neuroimage.usc.edu/pub/auditory_tutorial_T1.zipRight now I cannot use BrainSuite 16a on Windows 10 because the SVReg step is not working.
Francois
-
August 17, 2016 at 12:16 pm #298
Hi Francois,
I tried the volume you shared and it ran fine with all default parameters (no change in skull extraction or other parameters).
Can you try with default parameters and if it fails, can you please share the directory with all the outputs generated up to that point? I will try to see if I can figure out where things are diverging compared to my run on the same subject.
Thanks,
Best,
Anand
-
-
August 17, 2016 at 12:26 pm #299
-
August 17, 2016 at 2:08 pm #300
Hi Francois,
I am running SVReg sequence on the data that you shared and it is running fine. Do you have full Matlab or MCR for 2015b installed at the default location? Perhaps, one thing to try is that you have Matlab MCR for 2015b installed in default location. If you have any suggestions on how I can reproduce this error, that would be helpful. Do you have anything special in your Matlab/Windows installation (added third party library etc)? Anand -
August 17, 2016 at 2:41 pm #301
I have both the MCR 2015b and MATLAB 2015b installed on my computer.
None of them are in “standard installation path”.
I do not have any additional libraries saved in my Matlab path.How do you define from BrainSuite what is the base folder of Matlab?
NB: You should never be relying on “standard path”, this cannot be deployed properly.
On Windows systems, you can get the installation folders from the following registry keys:
HKLM\SOFTWARE\MathWorks\MATLAB\9.0
HKLM\SOFTWARE\MathWorks\MATLAB Runtime\9.0
HKLM\SOFTWARE\Wow6432Node\MathWorks\MATLAB\9.0
HKLM\SOFTWARE\Wow6432Node\MathWorks\MATLAB Runtime\9.0
Check the script brainstorm3/bin/R2015b/brainstorm3.bat for example on how to get these values from a shell script.-
August 17, 2016 at 4:31 pm #302
Hi Francois,
Thanks, if you are using windows system, then the compiled exe will figure out the path for MCR or Matlab runtime automatically, based on registry keys I think, but for Linux and Mac, one has to do some modifications to the svreg script.
I have never seen this error before and cann’t reproduce it here, so not sure how to proceed. I am trying it on another computer. If you have any suggestion on how to reproduce it here it would be helpful. Since the error came from settings of image processing toolbox, did you have any change in setting for that toolbox?
Anand
-
-
August 17, 2016 at 4:39 pm #303
I don’t have the signal processing toolbox in my MATLAB installation, I guess this is the source of the error.
In this case, I guess I should specify that I want to use the MCR instead of the full Matlab installation.
How can I change the Matlab base path that is used in BrainSuite? -
August 17, 2016 at 5:16 pm #304
According to the following link, it should not matter, but it also says the following.
Windows
To run deployed MATLAB code against MATLAB Runtime install, mcr_root\ver\runtime\win32|win64 must appear on your system path before matlabroot\runtime\win32|win64.
If mcr_root\ver\runtime\arch appears first on the compiled application path, the application uses the files in the MATLAB Runtime install area.
If matlabroot\runtime\arch appears first on the compiled application path, the application uses the files in the MATLAB installation area.It is worth trying to set environment PATH variable such that MCR appears before Matlab.
-
August 18, 2016 at 9:16 am #306
I exchanged the order of the two folders (MATLAB and MCR) in the system path and it looks like it’s working!
Thanks -
August 18, 2016 at 9:47 am #307
However, the results are better with the default parameters than with ‘Edge Constant=0.63’ and ‘Erosion Size=2’. If we do what you recommend on this MRI, there is a part of the frontal lobes that is not classified as brain and that is not included in the final cortex surfaces.
Can you please double-check this and then delete the comment you added again on the Brainstorm tutorial?
Thanks
Francois-
August 18, 2016 at 11:01 am #308
Hi Francois,
Yes, the default parameters work better for this volume and I deleted the comment about parameter change.
Thanks,
Anand
-
-
-
AuthorPosts
- You must be logged in to reply to this topic.