CTorch: PyTorch-Compatible GPU-Accelerated Auto-Differentiable Projector Toolbox for Computed Tomography

TL;DR

CTorch is introduced, a PyTorch-compatible, GPU-accelerated, and auto-differentiable projector toolbox designed to handle various CT geometries with configurable projector algorithms, with potential applications in accurate CT simulations, efficient iterative reconstruction, and advanced deep-learning-based CT reconstruction.

Abstract

This work introduces CTorch, a PyTorch-compatible, GPU-accelerated, and auto-differentiable projector toolbox designed to handle various CT geometries with configurable projector algorithms. CTorch provides flexible scanner geometry definition, supporting 2D fan-beam, 3D circular cone-beam, and 3D non-circular cone-beam geometries. Each geometry allows view-specific definitions to accommodate variations during scanning. Both flat- and curved-detector models may be specified to accommodate various clinical devices. CTorch implements four projector algorithms: voxel-driven, ray-driven, distance-driven (DD), and separable footprint (SF), allowing users to balance accuracy and computational efficiency based on their needs. All the projectors are primarily built using CUDA C for GPU acceleration, then compiled as Python-callable functions, and wrapped as PyTorch network module. This design allows direct use of PyTorch tensors, enabling seamless integration into PyTorch's auto-differentiation framework. These features make CTorch an flexible and efficient tool for CT imaging research, with potential applications in accurate CT simulations, efficient iterative reconstruction, and advanced deep-learning-based CT reconstruction.

Figures (8)

• Figure 1: Overall framework of the CTorch toolbox.
• Figure 2: Schematic diagram of the 2D and 3D circular scan geometries and associated parameters.
• Figure 3: Illustration of forward projection of a single voxel using different projector algorithms.
• Figure 4: FBP reconstruction using the top-layer projections from a dual-layer CBCT system wiht different projector algorithms.
• Figure 5: FBP reconstruction using the bottom-layer projections from a dual-layer CBCT system using different geometry definition. The non-circular geometry add a $0.4\degree$ in-plane rotation, effectively mitigating the edge blurring in the recon with circular geometry.