itk_convert MINC to NRRD metadata issues

[gdevenyi]

Tracking BIC-MNI/minc-toolkit-v2#70

[gdevenyi]

The issue appears to be at https://github.com/BIC-MNI/EZminc/blob/ITK4/tools/itk_convert.cpp#L131-L134 where the casting is failing and the resulting converion is to "0". Determinined by:

        std::cerr << itk::ExposeMetaData<double_vecto>( dict, "acquisition:bvalues",bvalues) <<std::endl;
        std::cerr << itk::ExposeMetaData<double_vector>( dict, "acquisition:direction_x",direction_x) <<std::endl;
        std::cerr << itk::ExposeMetaData<double_vector>( dict, "acquisition:direction_y",direction_y) <<std::endl;
        std::cerr << itk::ExposeMetaData<double_vector>( dict, "acquisition:direction_z",direction_z) <<std::endl;

Results in

0
0
0
0

Taking a look at the metadata with --show-meta

acquisition:b_value type: N3itk14MetaDataObjectIdEE
acquisition:bvalues type: N3itk14MetaDataObjectINS_5ArrayIdEEEE
acquisition:direction_x type: N3itk14MetaDataObjectINS_5ArrayIdEEEE
acquisition:direction_y type: N3itk14MetaDataObjectINS_5ArrayIdEEEE
acquisition:direction_z type: N3itk14MetaDataObjectINS_5ArrayIdEEEE

Not sure yet where that comes from. I think this indicates that the MINCIO isn't converting properly?

[vfonov]

two more examples (on BIC)
/scratch/scratch1/itk_convert/phantom_963271_MNI_SCAN1_20101125_dwi25_001.mnc
/scratch/scratch1/itk_convert/phantom_963271_MNI_SCAN1_20101125_dwi65_001.mnc

[vfonov]

The issue appears to be at https://github.com/BIC-MNI/EZminc/blob/ITK4/tools/itk_convert.cpp#L131-L134 where the casting is failing and the resulting converion is to "0". Determinined by:
...
Taking a look at the metadata with --show-meta

acquisition:b_value type: N3itk14MetaDataObjectIdEE
acquisition:bvalues type: N3itk14MetaDataObjectINS_5ArrayIdEEEE
acquisition:direction_x type: N3itk14MetaDataObjectINS_5ArrayIdEEEE
acquisition:direction_y type: N3itk14MetaDataObjectINS_5ArrayIdEEEE
acquisition:direction_z type: N3itk14MetaDataObjectINS_5ArrayIdEEEE

Not sure yet where that comes from. I think this indicates that the MINCIO isn't converting properly?
So, running
itk_convert --dti --minc-to-nrrd phantom_963271_MNI_SCAN1_20101125_dwi25_001.mnc phantom_963271_MNI_SCAN1_20101125_dwi25_001.nrrd

in .nrrd file , the acquisition:bvalues is empty, but in minc:
> mincheader phantom_963271_MNI_SCAN1_20101125_dwi25_001.mnc |grep acquisition:bvalues acquisition:bvalues = 0., 1000., 1000., 1000., 1000., 1000., 1000., 1000., 1000., 1000., 1000., 1000., 1000., 1000., 1000., 1000., 1000., 1000., 1000., 1000., 1000., 1000., 1000., 1000., 1000., 1000., 1000. ;

[gdevenyi]

Yup, sorry, I got that, the issue is that the code isn't properly converting/casting the minc-based metadata to a double_vector

[vfonov]

it's here: https://github.com/InsightSoftwareConsortium/ITK/blob/release-4.13/Modules/IO/MINC/src/itkMINCImageIO.cxx#L826

[vfonov]

Probably in https://github.com/BIC-MNI/EZminc/blob/ITK4/tools/itk_convert.cpp#L32
std::vector<double> should be replaced with Array<double>

[gdevenyi]

Fixed:

  Voxel->RAS x-form  : 
  Image Metadata: 
    DWMRI_b-value = 2000.000000
    DWMRI_gradient_0000 =  0.000000     0.000000     0.000000
    DWMRI_gradient_0001 =  1.000000    -0.000000     0.000000
    DWMRI_gradient_0002 =  0.705821    -0.515754     0.485607
    DWMRI_gradient_0003 =  0.288269    -0.615789     0.733284
    DWMRI_gradient_0004 = -0.079129    -0.581589     0.809625
    DWMRI_gradient_0005 =  0.194881    -0.308158     0.931161
    DWMRI_gradient_0006 = -0.160818    -0.247253     0.955512
    DWMRI_gradient_0007 =  0.204628    -0.958034     0.200743
    DWMRI_gradient_0008 = -0.601419    -0.109747    -0.791360
    DWMRI_gradient_0009 =  0.383061     0.822114     0.421180
    DWMRI_gradient_0010 = -0.702661     0.468698    -0.535341
    DWMRI_gradient_0011 = -0.189813    -0.696936    -0.691557
    DWMRI_gradient_0012 =  0.096749    -0.874742    -0.474833
    DWMRI_gradient_0013 = -0.955569    -0.294716    -0.005599
    DWMRI_gradient_0014 = -0.020564     0.879582     0.475303
    DWMRI_gradient_0015 =  0.069069    -0.834354     0.546885
    DWMRI_gradient_0016 = -0.318025    -0.947625    -0.029463
    DWMRI_gradient_0017 = -0.223853     0.674444     0.703573
    DWMRI_gradient_0018 = -0.316472    -0.750523     0.580138
    DWMRI_gradient_0019 = -0.418361     0.369888     0.829552
    DWMRI_gradient_0020 =  0.440707    -0.018973    -0.897450
    DWMRI_gradient_0021 = -0.997649     0.055741     0.039859
    DWMRI_gradient_0022 = -0.514338    -0.817170     0.260175
    DWMRI_gradient_0023 =  0.925726     0.290788    -0.241812
    DWMRI_gradient_0024 = -0.779804    -0.614015     0.122033
    DWMRI_gradient_0025 = -0.697030     0.683266     0.217479
    DWMRI_gradient_0026 =  0.176882     0.676630     0.714761
    DWMRI_gradient_0027 =  0.385588     0.607799    -0.694192
    DWMRI_gradient_0028 =  0.527175     0.565491     0.634276
    DWMRI_gradient_0029 = -0.732291    -0.211328     0.647372
    DWMRI_gradient_0030 =  0.794780     0.414608     0.443199
    DWMRI_gradient_0031 =  0.920259     0.377398     0.103415

[gdevenyi]

Fixed by #12