Welcome to dicom2nifti’s documentation!¶

Contents:

dicom2nifti¶

Python library for converting dicom files to nifti

API documentation:
Author:	Arne Brys
Organization:	icometrix
Repository:	https://github.com/icometrix/dicom2nifti
	http://dicom2nifti.readthedocs.io/en/latest

Using dicom2nifti¶

Installation¶

pip install dicom2nifti

Updating¶

pip install dicom2nifti --upgrade

Usage¶

Command line¶

dicom2nifti [-h] [-G] [-r] [-o RESAMPLE_ORDER] [-p RESAMPLE_PADDING] [-M] [-C] [-R] input_directory output_directory

for more information

dicom2nifti -h

From python¶

Converting a directory with dicom files to nifti files

import dicom2nifti

dicom2nifti.convert_directory(dicom_directory, output_folder)

Converting a directory with only 1 series to 1 nifti file

import dicom2nifti

dicom2nifti.dicom_series_to_nifti(original_dicom_directory, output_file, reorient_nifti=True)

Supported data¶

Most anatomical data for CT and MR should be supported as long as they are in classical dicom files.

Try avoiding “Implicit VR Endian” if possible as this makes converting non anatomical (i.e. DTI, fMRI, …) much more difficult.

There is some vendor specific support, more specifically for 4D imaging like fMRI and DTI/DKI

Gantry tilted CT¶

By default support for gantry tilted ct is disabled as we validate image orthogonality. You can explicitly allow gantry tilted data by disabling this validation.

Standard this will result in a nifti file where the gantry tilt is captured by the affine matrix. We also provide the option to resample the data to an orthogonal nifti. For this resampling we use scipy.ndimage.interpolation.affine_transform. You should configure the padding value and spline interpolation order

Command line:

dicom2nifti -G -r -o 1 -p -1000 input_directory output_directory

Python code:

import dicom2nifti
import dicom2nifti.settings as settings

settings.disable_validate_orthogonal()
settings.enable_resampling()
settings.set_resample_spline_interpolation_order(1)
settings.set_resample_padding(-1000)

dicom2nifti.convert_directory(dicom_directory, output_folder)

GE MR¶

Anatomical data should all be support. 4D images like fMRI and DTI/DKI are supported.

Siemens MR¶

Anatomical data should all be support. 4D images like fMRI and DTI/DKI are supported.

Philips MR¶

For classic dicom files dicom2nifti support anatomical. For classic dicom files 4D images like fMRI and DTI/DKI are supported.

For “Philips Enhanced Dicom” there is no support for “Implicit VR Endian” transfer syntax. For the others we support anatomical and 4D images like fMRI and DTI/DKI.

Hitachi MR¶

Anatomical data should all be support. 4D images like fMRI and DTI/DKI are NOT supported. Anyone willing to share DTI and/or fMRI dicom form Hitachi scanners please contact us.

Unsupported data¶

If you encounter unsupported data you can help the development of dicom2nifti by providing a dataset. This dataset should be anonymised (but leave as much of the private fields as possible).

dicom2nifti package¶

Submodules¶

dicom2nifti.common module¶

dicom2nifti

@author: abrys

dicom2nifti.common.apply_scaling(data, dicom_headers)[source]¶

Rescale the data based on the RescaleSlope and RescaleOffset Based on the scaling from pydicomseries

Parameters:	dicom_headers – dicom headers to use to retreive the scaling factors data – the input data

dicom2nifti.common.create_affine(sorted_dicoms)[source]¶

Function to generate the affine matrix for a dicom series This method was based on (http://nipy.org/nibabel/dicom/dicom_orientation.html)

Parameters:	sorted_dicoms – list with sorted dicom files

dicom2nifti.common.do_scaling(data, rescale_slope, rescale_intercept, private_scale_slope=1.0, private_scale_intercept=0.0)[source]¶

dicom2nifti.common.get_fd_array_value(tag, count)[source]¶