Quickstart#

This page projects an image, applies the matched adjoint and computes a gradient with autograd. Install diffct first; see Installation.

CUDA

import torch
from diffct import Projector, circular_trajectory_2d_parallel

trajectory = circular_trajectory_2d_parallel(180, device="cpu")
A = Projector(trajectory, (64, 64), 96, beam="parallel",
              detector_spacing=1.0, voxel_spacing=1.0)
image = torch.zeros(A.volume_shape, device="cuda", dtype=torch.float32)
image[20:44, 20:44] = 1.0
measurements = A.project(image)        # (180, 96)
adjoint = A.backproject(measurements)  # (64, 64), not an inverse

estimate = torch.zeros_like(image, requires_grad=True)
loss = 0.5 * (A.project(estimate) - measurements).square().sum()
loss.backward()                       # A^T (A estimate - measurements)
print(measurements.shape, estimate.grad.shape)

CPU (PyTorch backend)

import torch
from diffct import Projector, circular_trajectory_2d_parallel

trajectory = circular_trajectory_2d_parallel(180, device="cpu")
A = Projector(trajectory, (64, 64), 96, beam="parallel",
              detector_spacing=1.0, voxel_spacing=1.0, backend="torch")
image = torch.zeros(A.volume_shape, dtype=torch.float32)
image[20:44, 20:44] = 1.0
measurements = A.project(image)        # (180, 96)
adjoint = A.backproject(measurements)  # (64, 64), not an inverse

estimate = torch.zeros_like(image, requires_grad=True)
loss = 0.5 * (A.project(estimate) - measurements).square().sum()
loss.backward()                       # A^T (A estimate - measurements)
print(measurements.shape, estimate.grad.shape)

Projector defaults to beam="cone". Set beam="parallel" or beam="fan" for 2D data. Inputs must be floating-point CPU or CUDA tensors with the configured volume or sinogram shape in their trailing dimensions. Leading batch/channel dimensions are preserved. Outputs use the selected precision on the input device. Native CPU data streams through CUDA tiles; the PyTorch backend computes on its input device. See Precision, training and reconstruction and Large Volumes and Chunked Execution. Trajectory tensors can stay on CPU.

Next steps#

A custom or non-circular scan: see Trajectories and Geometry.
A curved detector: see Parameterized detector surfaces and run python examples/curved_detector.py from a source checkout.
An iterative reconstruction: run python examples/iterative_reconstruction.py --trajectory helical. See Examples.
Several GPUs or nodes: see Multiple GPUs and Nodes.
Code written for diffct 1.x: see Migrating from diffct 1.x to 2.0.
Gradient-based reconstruction with autograd and an optimizer: see Gradient-Based Iterative Reconstruction.

Troubleshooting#

Shape error: use (D, H, W) for cone volumes and (views, U, V) for cone sinograms. In distributed mode, use A.projection_shape and A.view_slice rather than the global view count.