diffct: Differentiable Computed Tomography Operators

A high-performance, CUDA-accelerated library for circular orbit CT reconstruction with end-to-end differentiable operators, enabling advanced optimization and deep learning integration in medical imaging and scientific computing.

Key Features

High Performance: CUDA-accelerated projection and backprojection operations with optimized memory management
Fully Differentiable: End-to-end gradient propagation through all CT operations for seamless deep learning integration
Multiple Geometries: Support for 2D parallel-beam, 2D fan-beam, and 3D cone-beam geometries
PyTorch Integration: Native PyTorch autograd support with custom CUDA kernels
Research Ready: Optimized for both analytical reconstruction (FBP/FDK) and iterative methods

Supported Geometries

Parallel Beam (2D): Traditional parallel-beam geometry for 2D CT reconstruction
Fan Beam (2D): Fan-beam geometry with configurable source-detector distances
Cone Beam (3D): Full 3D cone-beam geometry for volumetric reconstruction

Applications

Medical image reconstruction with deep learning enhancement
Physics-informed neural networks for CT imaging
Iterative reconstruction algorithms with learned priors
Multi-modal imaging research and development
Educational CT reconstruction demonstrations

Getting Started

Getting Started

Getting Started

User Guide

User Guide

Citation

If you use this library in your research, please cite:

@software{DiffCT2025,
  author       = {Yipeng Sun},
  title        = {DiffCT: Differentiable Computed Tomography
                 Reconstruction with CUDA},
  year         = 2025,
  publisher    = {Zenodo},
  doi          = {10.5281/zenodo.14999333},
  url          = {https://doi.org/10.5281/zenodo.14999333}
}

License

This project is licensed under the Apache 2.0 - see the LICENSE file for details.