pyronn-torch

pyronn-torch

This repository provides PyTorch bindings for PYRO-NN, a collection of back-rotatable projectors for CT reconstruction.

Feel free to cite our publication:

@article{PYRONN2019,
author = {Syben, Christopher and Michen, Markus and Stimpel, Bernhard and Seitz, Stephan and Ploner, Stefan and Maier, Andreas K.},
title = {Technical Note: PYRO-NN: Python reconstruction operators in neural networks},
year = {2019},
journal = {Medical Physics},
}

Installation

From PyPI:

pip install pyronn-torch

From this repository:

git clone --recurse-submodules --recursive https://github.com/theHamsta/pyronn-torch.git
cd pyronn-torch
pip install torch
pip install -e .

You can build a binary wheel using

python setup.py bdist_wheel

Usage

#ConeBeamProjector(volume_shape,
#                  volume_spacing,
#                  volume_origin,
#                  projection_shape,
#                  projection_spacing,
#                  projection_origin,
#                  projection_matrices)
projector = pyronn_torch.ConeBeamProjector(
    (128, 128, 128),
    (2.0, 2.0, 2.0),
    (-127.5, -127.5, -127.5),
    (2, 480, 620),
    [1.0, 1.0],
    (0, 0),
    np.array([[[-3.10e+2, -1.20e+03,  0.00e+00,  1.86e+5],
               [-2.40e+2,  0.00e+00,  1.20e+03,  1.44e+5],
               [-1.00e+00,  0.00e+00,  0.00e+00,  6.00e+2]],
              [[-2.89009888e+2, -1.20522754e+3, -1.02473585e-13,
                1.86000000e+5],
               [-2.39963440e+2, -4.18857765e+0,  1.20000000e+3,
                1.44000000e+5],
               [-9.99847710e-01, -1.74524058e-2,  0.00000000e+0,
                6.00000000e+2]]]) # two projection matrices
)
projection = projector.new_projection_tensor(requires_grad=True)

projection += 1.
result = projector.project_backward(projection, use_texture=with_texture)

assert projection.requires_grad
assert result.requires_grad

loss = result.mean()
loss.backward()

Or easier with PyCONRAD (pip install pyconrad)

projector = pyronn_torch.ConeBeamProjector.from_conrad_config()

The configuration can then be done using CONRAD (startable using conrad from command line)

Release files for pyronn-torch 0.0.4

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for pyronn-torch 0.0.4
File	Size	Uploaded
pyronn-torch-0.0.4.tar.gz	30.6 kB	Feb 27, 2020	Details

Release files / pyronn-torch-0.0.4.tar.gz

Download URL	pyronn-torch-0.0.4.tar.gz
Size	30.6 kB
Tags	Source
SHA-256 checksum How to use checksums	`3e379345a985452b64c52aac50e015494c048a5758d0fc00aa419b06920b04c1`
BLAKE2b-256 checksum How to use checksums	`30beac305b716a6a330eac1c098c8ac906ab6653e55690208f77e7b0a0d4a987`
Upload date	Feb 27, 2020
Uploaded using Trusted Publishing? What is trusted publishing?	No
Uploaded via	`twine/1.13.0 pkginfo/1.5.0.1 requests/2.21.0 setuptools/45.1.0 requests-toolbelt/0.8.0 tqdm/4.40.2 CPython/3.7.5`