MBIR Cone

Python Package for Cone Beam Computed Tomography reconstruction. Full documentation is available at https://mbircone.readthedocs.io

Warning: This is a pre-release version of code that is still under development.

Distribution Statement

Distribution Statement A. Approved for public release: distribution unlimited (88ABW-2020-0895).

If you publish results based on this code, please cite the following paper:

Thilo Balke, Soumendu Majee, Gregery T. Buzzard, Scott Poveromo, Patrick Howard, Michael A. Groeber, John McClure, Charles A. Bouman "Separable Models for cone-beam MBIR Reconstruction," Proceedings of the IS&T International Symposium on Electronic Imaging, Computational Imaging XVI, pp. 181-1 to 181-7, 2018.

For other OpenMBIR packages see: https://github.com/cabouman/OpenMBIR-Index

Installation

In order to install the package, clone the repository to your computer, and run the bash script named run_clean_install. The script will create and activate a conda envirnoment named mbircone, install the package and all its requirements, and build the documentation.

The information below provides additional detail.

1. Clone Repository and enter

git clone https://github.com/cabouman/mbircone.git
cd mbircone

2. Create conda environment

conda create -n mbircone python=3.8

3. Activate conda environment

conda activate mbircone

4. Install requirements

pip install -r requirements.txt

5. Ensure GCC is installed per instructions here: https://svmbir.readthedocs.io/en/latest/install.html

6. Install package

pip install .

Run demos

1. Install demo requirements

cd demo
pip install -r requirements_demo.txt

2. Run basic demo

python demo_3D_shepp_logan.py

3. Run MACE demo

python demo_mace3D.py

4. (Optional) Run NSI dataset demo. (Note: this is a long demo with an estimated run time of 30-60 minutes!)

python demo_nsi_preprocess.py

5. Result visualization:

Please go to demo/output/ to look at phantom, sinogram, and reconstruction images

6. In the case where exceptions occur when downloading data, please check your internet connection. If you replaced the default url with the url of your own dataset, please make sure that the url is correct, and points to a public webpage.

Build documentation in local folder

1. Install docs requirements

cd docs
pip install -r requirements.txt

2. Build documentation

MBIRCONE_BUILD_DOCS=true make html

3. Open documentation

cd build/html
open(double clicks) index.html