Dicomifier is a set of tools to convert Bruker data to DICOM files, and DICOM files to NIfTI. It retains meta-data (e.g. MR parameters such as echo time or subject parameters such as weight or height) throughout the conversion process, storing them in JSON format, allowing easy access in multiple environments (command-line, C++, MATLAB, Python). The meta-data from Bruker is aligned on the DICOM dictionary for unified human and animal processing pipelines.
Dicomifier builds and runs on:
- Linux (Debian 8 and 9, Ubuntu 16.04 and 18.04, CentOS 7).
- OS X
A Docker image is also available.
Unofficial packages are provided for the Debian and Ubuntu listed above. After following the instructions to set up the unofficial repository, install the dicomifier-cli package with the usual tools provided by your distribution, e.g.:
- Debian/Ubuntu:
sudo apt-get install dicomifier-cli - CentOS:
sudo yum install dicomifier-cli
If your distribution has no pre-compiled package available, you can either compile it from source (cf. documentation for Linux and macOS) or ask for packaging.
Bruker-to-DICOM conversion is performed by bruker2dicom. This tool can use either uncompressed Bruker directories or PvDataset files and has been field-tested on different acquisitions types (anatomical, functional, diffusion) and different versions of Paravision (4, 5 and 6).
To list the series, use bruker2dicom list:
computer:~ user$ bruker2dicom list lb_140721.Bq1/
1:1 - 1_Localizer (USER_MODE)
2:1 - T2_TurboRARE (USER_MODE)
3:1 - T2_TurboRARE (USER_MODE)
4:1 - T2_TurboRARE_3D (USER_MODE)
5:1 - (none) (USER_MODE)
6:1 - T2_TurboRARE (USER_MODE)
7:1 - T2_TurboRARE_3D (USER_MODE)
8:1 - T2_TurboRARE_3D (USER_MODE)
9:1 - T2map_MSME (USER_MODE)
10:1 - T2map_MSME (USER_MODE)
11:1 - T1map_RARE (USER_MODE)
11:2 - (none) (none)To convert to DICOM, run bruker2dicom convert:
computer:~ user$ bruker2dicom convert lb_140721.Bq1/ dicom/
WARNING - dicomifier: Skipping 11:2 - (none) (none): type is DERIVED_ISAA single series/reconstruction can be targetted with the -r option: running bruker2dicom convert -r 9:1 lb_140721.Bq1/ dicom/ will only convert the first T2map_MSME series. More information about the conversion process can be displayed using the -v option with different verbosity levels (warning, the default, info or debug):
computer:~ user$ bruker2dicom convert -r 8:1 -v debug lb_140721.Bq1/ dicom
INFO - dicomifier: Converting 8:1
INFO - dicomifier: Found 8:1 - T2_TurboRARE_3D (USER_MODE)Note that the tool will warn you about series that it cannot convert: some Bruker derived data (e.g. diffusion tensor) have no public specification available and are skipped.
The DICOM data will be stored with a subject/study/series/reconstruction hierarchy. This way, multiple exams of the same subject may be stored in the same directory. An index of the DICOM files (i.e. a DICOMDIR file) can be created at the top-level of the destination directory by using the -d option.
DICOM-to-NIfTI conversion is performed by dicom2nifti. Its inputs are a set of either directories containing DICOM files or DICOMDIR files, and the destination directory of NIfTI data:
computer:~ user$ dicom2nifti dicom/ nifti/`The NIfTI data will be stored with a subject/study/series. Each series directory will contain all the reconstructions of the series, and each reconstruction will have two files: the NIfTI file itself, and the meta-data associated with it, in the JSON format. The subject, study and series directories will be named according to the information included in the DICOM file:
computer:~ user$ tree -L 3 nifti/
nifti/
└── fantome-nouille
└── 7_plateforme^18062016
├── 1114113_T2_TurboRARE
├── 1245185_B0Map-ADJ_B0MAP
├── 1966081_Perfusion_FAIR_RARE
├── 589825_T2map_MSME
├── 655361_T2map_MSME
├── 65537_1_Localizer
└── 917505_Perfusion_FAIR_RARECompressed NIfTI files (nii.gz) can be created by passing the -z option, and more information about the conversion process can be displayed using the -v option with different verbosity levels (warning, the default, info or debug):
computer:~ user$ dicom2nifti -z -v debug dicom/ nifti/
INFO - dicomifier: Splitting 62 DICOM files in series
INFO - dicomifier: 1 series found
INFO - dicomifier: Reading 62 DICOM files
INFO - dicomifier: Found 2 stacks
INFO - dicomifier: Converting 7-plateforme^18062016 / 589825-T2map_MSME
DEBUG - dicomifier: Converting 7-plateforme^18062016 / 589825-T2map_MSME (stack 1/2)
INFO - dicomifier: dtype deduced to be: <type 'numpy.float32'>
DEBUG - dicomifier: Converting 7-plateforme^18062016 / 589825-T2map_MSME (stack 2/2)
INFO - dicomifier: dtype deduced to be: <type 'numpy.float32'>
INFO - dicomifier: dtype deduced to be: <type 'numpy.float32'>
INFO - dicomifier: Merging 2 stacks
computer:~ user$ ls -l nifti/*/*/589825_T2map_MSME/
total 9944
-rw-rw---- 1 user group 2.4K 23 déc 12:29 1.json
-rw-rw---- 1 user group 4.9M 23 déc 12:29 1.nii.gzThe meta-data is stored in the JSON file as a dictionary where the entries are the DICOM names, and the values are always stored as arrays, even for single-valued entries. Since the number of element in a value can vary between exams, this choice yields a uniform data storage and data access. For example, the echo time is stored in the EchoTime field and can be retrieved as such:
- Command-line, using jq:
or, for series with multiple echoes:
computer:~/nifti user$ jq '.EchoTime[0]' 1.json 3
computer:~/nifti user$ jq '.EchoTime[][0]' 1.json 7.699999809265137 15.39999961853027 23.10000038146973 30.79999923706055 38.5 46.20000076293945 53.90000152587891
- Python, using the standard module
import json meta_data = json.load(open("1.json")) # Single echo: echo_time = meta_data["EchoTime"][0] # 4D series, multiple echoes: echo_times = [x[0] for x in meta_data["EchoTime"]]
The complete specification of this JSON format, including 4D or higher series, is available in the wiki.