Deformation inversion layer is a neural network layer for inverting deformation fields develped as part of SITReg, a deep learning intra-modality image registration arhitecture fulfilling strict symmetry properties.
Install using pip by running the command
pip install deformation-inversion-layer
Python 3.8+PyTorch 1.10+
For a quick start tutorial, see quick_start.ipynb. For API reference, go to https://honkamj.github.io/deformation-inversion-layer/.
For SITReg implementation, see repository SITReg.
If you use deformation inversion layer, or other parts of the repository, please cite (see bibtex):
- SITReg: Multi-resolution architecture for symmetric, inverse consistent, and topology preserving image registration
Joel Honkamaa, Pekka Marttinen
The Journal of Machine Learning for Biomedical Imaging (MELBA) (10.59275/j.melba.2024-276b)
Tutorial by Zico Kolter, David Duvenaud, and Matt Johnson was very helpful in implementing the layer.
Deformation inversion layer and SITReg are released under the MIT license.