CTLIB

A lib of CT projector and back-projector based on PyTorch

Coded with distance driven method [1, 2]

If you use the code, please cite our work

@article{xia2021magic,
  title={MAGIC: Manifold and Graph Integrative Convolutional Network for Low-Dose CT Reconstruction},
  author={Xia, Wenjun and Lu, Zexin and Huang, Yongqiang and Shi, Zuoqiang and Liu, Yan and Chen, Hu and Chen, Yang and Zhou, Jiliu and Zhang, Yi},
  journal={IEEE Transactions on Medical Imaging},
  year={2021},
  publisher={IEEE}
}

Installation

The following is the step-by-step instruction to install this lib using conda

Create Conda environment

conda create -n ctlib

Enter the enviroment and install pytorch

conda activate ctlib
conda install pytorch torchvision torchaudio pytorch-cuda=11.8 -c pytorch -c nvidia

Notice: Usually the cudatoolkit installed from nvidia channel will provide the complier. But if you install the previous pytorch version using earlier cudatoolkit which is not from nvidia channel, the cudatoolkit may not include the complier. The later installation will report the error that can't find nvcc. If so, you need install the lib of cudatoolkit-dev as follows:

conda install cudatoolkit-dev -c conda-forge

Move to the the directory of this lib, and then compile and install

python setup.py install

API

projection(image, options): Projector of CT

projection_t(projection, options): Transpose of projector

backprojection_t(image, options): Transpose of backprojector

backprojection(projection, options): Backprojector of CT

fbp(projection, options): FBP with RL filter

laplacian(input, k): Computation of adjancency matrix in [3]

image: 4D torch tensor, B x 1 x H x W,

projection: 4D torch tensor, B x 1 x V x D, V is the total number of scanning views, D is the total number of detector bins

options: 12D torch vector for fan beam and 9D torch vector for parallel beam, scanning geometry parameters, including

views: Number of scanning views

dets: Number of detector bins

width and height: Spatial resolution of images

dImg: Physical length of a pixel

dDet: Interval between two adjacent detector bins, especially, interval is rad for equal angle fan beam

Ang0: Starting angle

dAng: Interval between two adjacent scanning views: rad

s2r: The distance between x-ray source and rotation center, not needed in parallel beam

d2r: The distance between detector and roration center, not needed in parallel beam

binshift: The shift of the detector

scan_type: 0 is equal distance fan beam, 1 is euql angle fan beam and 2 is parallel beam

[1] B. De Man and S. Basu, “Distance-driven projection and backprojection,” in IEEE Nucl. Sci. Symp. Conf. Record, vol. 3, 2002, pp. 1477–80.

[2] B. De Man and S. Basu, “Distance-driven projection and backprojection in three dimensions,” Phys. Med. Biol., vol. 49, no. 11, p. 2463, 2004.

[3] Xia, W., Lu, Z., Huang, Y., Shi, Z., Liu, Y., Chen, H., ... & Zhang, Y. (2021). MAGIC: Manifold and Graph Integrative Convolutional Network for Low-Dose CT Reconstruction. IEEE Transactions on Medical Imaging.