Paper: Multi-Scale Diffusion Model for Ultra-Sparse View CT Reconstruction
Authors: Junyan Zhang, Mengxiao Geng, Pinhuang Tan, Yi Liu, Zhili Liu, Bin Huang, Qiegen Liu
https://iopscience.iop.org/article/10.1088/1361-6560/ae2fa7/meta
Physics in Medicine & Biology, Volume 71, Number 1.

Abstract: Computed Tomography (CT) technology reduces radiation exposure to the human body through sparse sampling, but fewer sampling angles pose challenges for image reconstruction. When the projection angles are significantly reduced, the quality of image reconstruc-tion deteriorates. To improve the quality of image reconstruction under sparse angles, an ultra-sparse view CT reconstruction method utilizing multi-scale diffusion models is proposed. This method aims to focus on the global distribution of information while facilitating the reconstruction of local image features in sparse views. Specifically, the proposed model ingeniously combines information from both comprehensive sampling and selective sparse sampling techniques. By precisely adjusting the diffusion model, diverse noise distribu-tions are extracted, enhancing the understanding of the overall image structure and assisting the fully sampled model in recovering image information more effectively. By leveraging the inherent correlations within the projection data, an equidistant mask is designed according to the principles of CT imaging, allowing the model to focus attention more efficiently. Experimental results demonstrate that the multi-scale model approach significantly improves image reconstruction quality under ultra-sparse views and exhibits good gener-alization across multiple datasets.

Index Terms—Computed tomography, multi-diffusion model, ultra-sparse view reconstruction, sinogram domain.

(a) The reference image, (b) FBP, (c) U-Net, (d) FBPConvNet, (e) GMSD, (f) MSDiff.

(a) The reference image, (b) FBP, (c) U-Net, (d) FBPConvNet, (e) GMSD, (f) MSDiff.

Full-view Diffusion Model (FDM)

python main_720.py --config=aapm_sin_ncsnpp_720.py --workdir=exp_fd --mode=train --eval_folder=result

Sparse-view Diffusion Model (SDM)

python main_120.py --config=aapm_sin_ncsnpp_120.py --workdir=exp_sd --mode=train --eval_folder=result

python PCsampling_demo.py

• Generative Modeling in Sinogram Domain for Sparse-view CT Reconstruction
  [Paper] [Code]

• One Sample Diffusion Model in Projection Domain for Low-Dose CT Imaging
  [Paper] [Code]

• Iterative Reconstruction for Low-Dose CT using Deep Gradient Priors of Generative Model
  [Paper] [Code] [PPT]

• REDAEP: Robust and Enhanced Denoising Autoencoding Prior for Sparse-View CT Reconstruction
  [Paper] [Code] [PPT]

• Wavelet-improved score-based generative model for medical imaging
  [Paper]

• 基于深度能量模型的低剂量CT重建
  [Paper] [Code]

• Stage-by-stage Wavelet Optimization Refinement Diffusion Model for Sparse-view CT Reconstruction
  [Paper] [Code]

• Dual-Domain Collaborative Diffusion Sampling for Multi-Source Stationary Computed Tomography Reconstruction
  [Paper] [Code]

• Low-rank Angular Prior Guided Multi-diffusion Model for Few-shot Low-dose CT Reconstruction
  [Paper] [Code]

• Physics-informed DeepCT: Sinogram Wavelet Decomposition Meets Masked Diffusion
  [Paper] [Code]

• MSDiff: Multi-Scale Diffusion Model for Ultra-Sparse View CT Reconstruction
  [Paper] [Code]

• Ordered-subsets Multi-diffusion Model for Sparse-view CT Reconstruction
  [Paper]

• Virtual-mask Informed Prior for Sparse-view Dual-Energy CT Reconstruction
  [Paper] [Code]

• Raw_data_generation [Code]

• PRO: Projection Domain Synthesis for CT Imaging [Paper] [Code]

• UniSino: Physics-Driven Foundational Model for Universal CT Sinogram Standardization[Paper] [Code]

• Diffusion Models for Medical Imaging [Paper] [Code] [PPT]