CUDA-accelerated Python utilities for high-throughput neuroimage processing and analysis


Keywords
PET, MR, processing, analysis, cuda, gpu, medical-imaging, python
License
Apache-2.0
Install
pip install nimpa==2.6.3

Documentation

NIMPA: Neuro and NiftyPET Image Processing and Analysis

Docs Version Downloads Py-Versions DOI Licence Tests

NIMPA is a stand-alone Python sub-package of NiftyPET, dedicated to high-throughput processing and analysis of brain images, particularly those, which are acquired using positron emission tomography (PET) and magnetic resonance (MR). Although, it is an essential part of the NiftyPET package for seamless PET image reconstruction, NIMPA is equally well suited for independent image processing, including image trimming, upsampling and partial volume correction (PVC).

Trimming is performed in order to reduce the unused image voxels in brain imaging, when using whole body PET scanners, for which only some part of the field of view (FOV) is used.

The upsampling is needed for more accurate extraction (sampling) of PET data using regions of interest (ROI), obtained using parcellation of the corresponding T1w MR image, usually of higher image resolution.

PVC is needed to correct for the spill-in and spill-out of PET signal from defined ROIs (specific for any given application).

Documentation with installation manual and tutorials: https://niftypet.readthedocs.io/

Quick Install

Note that it's recommended (but not required) to use conda.

# cross-platform install
conda install -c conda-forge python=3 \
  ipykernel numpy scipy scikit-image matplotlib ipywidgets dipy nibabel pydicom
pip install "nimpa>=2"

For optional dcm2niix (image conversion from DICOM to NIfTI) and/or niftyreg (image registration) support, simply install them separately (pip install dcm2niix niftyreg).

External CMake Projects

The raw C/CUDA libraries may be included in external projects using cmake. Simply build the project and use find_package(NiftyPETnimpa).

# print installation directory (after `pip install nimpa`)...
python -c "from niftypet.nimpa import cmake_prefix; print(cmake_prefix)"

# ... or build & install directly with cmake
mkdir build && cd build
cmake ../niftypet && cmake --build . && cmake --install . --prefix /my/install/dir

At this point any external project may include NIMPA as follows (Once setting -DCMAKE_PREFIX_DIR=<installation prefix from above>):

cmake_minimum_required(VERSION 3.3 FATAL_ERROR)
project(myproj)
find_package(NiftyPETnimpa COMPONENTS improc REQUIRED)
add_executable(myexe ...)
target_link_libraries(myexe PRIVATE NiftyPET::improc)

Licence

Licence DOI

Copyright 2018-21