DEnse Displacement Sampling - deformable image registration

This package provides Python wrapper around DEEDS, an efficient version for 3D discrete deformable image registration which is reaching the highest accuracy in several benchmarks [1][2] and serves as a good baseline for new solutions.

Referencing and citing

If you use this implementation or parts of it please cite:

"MRF-Based Deformable Registration and Ventilation Estimation of Lung CT." by Mattias P. Heinrich, M. Jenkinson, M. Brady and J.A. Schnabel IEEE Transactions on Medical Imaging 2013, Volume 32, Issue 7, July 2013, Pages 1239-1248 http://dx.doi.org/10.1109/TMI.2013.2246577

"Multi-modal Multi-Atlas Segmentation using Discrete Optimisation and Self-Similarities" by Mattias P. Heinrich, Oskar Maier and Heinz Handels VISCERAL Challenge@ ISBI, Pages 27-30 2015 http://ceur-ws.org/Vol-1390/visceralISBI15-4.pdf

and

"DEEDS Flow Field" by Alexis Coullomb and Douglas Shepherd 10.5281/zenodo.15366235

Installation

pip install git+https://github.com/wiktorowski211/deeds-registration

The build automatically detects if your CPU supports AVX2 instructions and uses them when available. You can force or disable AVX2 usage with the USE_AVX2 environment variable (set USE_AVX2=1 to force enable or USE_AVX2=0 to disable):

USE_AVX2=0 pip install git+https://github.com/wiktorowski211/deeds-registration

Usage

from deeds import registration
import SimpleITK as sitk

fixed = sitk.ReadImage(PATH)
moving = sitk.ReadImage(PATH)

moved = registration(fixed, moving)

Prerequesities

Input images must:

• have the same dimensions,
• be a SimpleITK image object.

Development

Build:

python setup.py build_ext --inplace

Test:

python -m unittest