PyEPRI: A CPU & GPU compatible Python package for Electron Paramagnetic Resonance Imaging
Résumé
This work presents the PyEPRI package, an open-source Python package for Electron Paramagnetic Resonance Imaging. The PyEPRI package implements low-level operators, like projection and backprojection, involved in Electron Paramagnetic Resonance (EPR) and also high-level advanced algorithms, like total variation based EPR image reconstruction, for end-users. The package is fully implemented in Python and provides both CPU and GPU computation capabilities, through the libraries Numpy, PyTorch and Cupy. This package comes with a detailed documentation, including precise mathematical definitions and many reproducible demonstration examples and tutorials, making it easy for users with no particular expertise on coding image processing algorithms to get started. This package is also highly modular and only relies on standard data types, as such, it can also be easily used by advanced users to develop new algorithms while benefiting from an optimized computing environment and some rigorously tested operators. The PyEPRI package has been developed by researchers with the hope that it will be useful to the EPR community.
Origine | Fichiers produits par l'(les) auteur(s) |
---|