SCI Seg3D Mailing List

Text archives Help


Re: [Seg3D] Shifting of the files


Chronological Thread 
  • From: Jess Tate <jess@sci.utah.edu>
  • To: Andjela Davidovic <andjela.davidovic@inria.fr>
  • Cc: scirun-users@sci.utah.edu, seg3d@sci.utah.edu, ayla <ayla@sci.utah.edu>
  • Subject: Re: [Seg3D] Shifting of the files
  • Date: Mon, 3 Aug 2015 22:47:09 -0600
Hi Andjela,

Different nrrd formats can be interpreted differently by different software.  That unfortunately could be true for this case with Seg3D and SCIrun, though it shouldn't be.  However, the error that you are seeing with the data size error is not because of this.  When using CalculateFieldData with multiple fields, the two fields must have the same number of data points, but the module doesn't care if the geometry is the same.  In your case it is because you have a more refined grid on one geometry than the other.  The solution in this case is to Use MapFieldDataOntoNodes/Elements.  Both fields will need to match node or element centering to get the number of data values to match.  Also, this will depend on getting the geometries to match. Try loading them both into SCIRun and using reportFieldInfo (i key in SCIRun5).  The center and size should match, if not, then we will need to figure out why not.

Alternatively, you can  resample either volume to match the other (ResampleRegularGrid in SCIrun or Resample tool in  Seg3D) and then use the method you talked about the mask the fibers.

Does that make sense?

cheers,
Jess




On Aug 3, 2015, at 10:18 PM, Andjela Davidovic <andjela.davidovic@inria.fr> wrote:

Hi,

It seems there is some problem with file import/export in Seg3D and SCIRun, with spacings and origins. 

So, my original files are of the following formats in .vtk:
MRI data: 256x256x512, spacing  0.05, 0.05, 0.05,  origin 0, 0, 0
Fibers files: 64x64x128, spacing 0.2, 0.2, 0.2, origin 0, 0, 0


Then I use Seg3d to segment my MRI data, and for some reason the output of the segmentation in .nrrd format becomes:
256x256x512, spacing  0.05, 0.05, 0.05,  origin  -0.025, -0.025, -0.025

I also use Seg3d to convert Fiber files in .nrrd format, and for some reason th output becomes
64x64x128, spacing 0.2, 0.2, 0.2, origin -0.1, -0.1, -0.1

This messes finally also with the size of the domain, and then SCIRun thinks that the data are not of the same size, and does not want to ComputeFieldData - in order to project fibers onto segmentation.

I tried to use Transform in Seg3d, to force origin to 0, 0, 0, and it appears to work at first. When I open the file in vi it shows origin at 0, 0, 0, but when I import file in SCIRun it seems to be shifted again and of the wrong size.

Any advice?

Andjela


--
Andjela Davidovic
PhD Student, CARMEN team
INRIA
CENTRE DE RECHERCHE
BORDEAUX - SUD-OUEST
200 avenue de la Vieille Tour
33 405 Talence Cedex
+33 (0)5 24 57 41 64



  • Re: [Seg3D] Shifting of the files, Jess Tate, 08/03/2015

Archive powered by MHonArc 2.6.18.

Top of page