From Sparse X-rays to 3D CT: Training-Free Reconstruction with Diffusion Priors
Updated
Updated · arxiv.org · Jul 15
From Sparse X-rays to 3D CT: Training-Free Reconstruction with Diffusion Priors
1 articles · Updated · arxiv.org · Jul 15
Summary
Researchers have introduced TF-PRDiT, a training-free framework that reconstructs 3D CT images from sparse X-ray data using a frozen diffusion prior.
The method enforces measurement consistency during sampling, allowing a single pretrained model to adapt to varying numbers of X-ray views without retraining.
TF-PRDiT achieves strong performance on multiple inverse medical problems, potentially streamlining 3D medical imaging workflows and reducing reliance on large supervised datasets.