This script enables converting Interfile files to DICOM file format and importing meta-data from external json file.
Prerequisites:
- Python (vers. 3.8.X or newer recommended)
- Libraries from requirements.txt
If you use Debian/Ubuntu/Linux Mint you can install python3 using 'apt-get install' command:
sudo apt-get install python3
To download the required modules, please run:
pip install -r requirements.txt
Where "requirements.txt" is in the repository.
Package provides an API for converting Interfile file into Dicom files. Interfile is specified by two files:
- header file (.hdr), that has CASToRv3.1 keyset
- image file (specified in the header)
To convert a file/files using a command line, run the main.py file, for example:
python3 main.py -i header.hdr -m metadata.json
Using a --help flag will show all available commands.
There are many applications to visualize images in DICOM files. Personally, I recommend using Amide:
To open a .dcm file use the FILE menu:
File > Import File(specify) > DICOM 3.0 via (X)MedCon.
Amide uses a very useful tool called (X)MedCon, which can be downloaded from project's website or system's repositories, and it can be used alone to open DICOM files (but Amide provides better visualization).
To install (X)MedCon on Debian/Ubuntu/Linux Mint from repositories:
sudo apt-get install xmedcon
In order to open DICOM file from Terminal, use command:
xmedcon -f name_of_file.dcm
(X)MedCon has many other options and supported formats, see project's documentation for details.
This section talks about previous binary2DICOM module, made by Rafał Masałek. In this version it's still partially used by the module, but it's deprecated and does not work correctly. It can still convert raw binary data into a dicom file, but it doesn't have mandatory tags.
Usage example: This is an example command used to convert Gate output to DICOM format. File with simulated data is called in this example collimator_20.
python3 binary2DICOM.py --meta meta.txt collimator_20 collimator_20-conv.dcm 256 256 256 -bpp 4 -int -bo little
- python3 binary2DICOM.py uses python3 to run the script
- --meta meta.txt specifies that external file with meta data is used; it is called 'meta.txt' and is located in the same directory as the binary2DICOM.py script
- collimator_20 name of the binary image file (input file for script)
- collimator_20-conv.dcm name of the output file, note the '.dcm' extension
- 256 256 256 X, Y and Z(no of frames) dimensions of encoded images (size in pixels)
- -bpp 4 defines that 4 bytes were used to encode every pixel of input data (if you don't know encoding type, then ask the maker of the input file)
- -int specifies that signed integers were used to encode input image (use '-uint' for unsigned integers)
- -bo little specifies the byte order used to encode binary image; supported are little/big
Not all parameters are necessary, optional ones have default values. You can get more information by using --help flag.
The script is capable of reading binary images encoded using signed or unsigned integers, 1/2/4/8 bytes per pixel. In the future it might be developed to manage file encoded using floating point numbers.
External file with metadata must have apropriate format (see meta.txt): You have to specify: group tag, element tag, VR and value. All these elements have to be separated using comma. String values need to be inside ' characters. Dot is used as a decimal delimiter. You can write comments using hash symbol (#) -- following characters are ignored.
The script can be used as a python module in external program. To use it you have to import binary2DICOM to your script:
from binary2DICOM import convert
Then you have to call convert() method, providing special dictionary as an argument. The dictionary must have the following structure:
{'meta': (meta_data_file_path, STRING),
'in_file': (input_data_file_path, STRING),
'out_file': (output_data_file_path, STRING),
'width': (INT),
'height': (INT),
'frames': (INT),
'is_signed': (BOOL),
'byte_order': ('little' OR 'big'),
'bytes_per_pix': (INT),
'is_float': (BOOL)}
Key-value pairs in the dictionary correspond to input parameters. You code should look like this:
convert({'meta': 'meta.txt', 'width': 256, 'is_signed': True, 'in_file': 'collimator_20', 'byte_order': 'little', 'height': 256, 'frames': 256, 'bytes_per_pix': 4, 'out_file': 'collimator_20-conv.dcm', 'is_float': False})
Of course you can store dictionary in a variable and pass this variable to the convert() function.