Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Volumetric Conditional Score-based Residual Diffusion Model for PET/MR Denoising

Siyeop Yoon, Rui Hu, Yuang Wang, Matthew Tivnan, Young-don Son, Dufan Wu, Xiang Li, Kyungsang Kim, Quanzheng Li

TL;DR

The Conditional Score-based Residual Diffusion (CSRD) model is presented, which significantly lowers computational demands and expedites the denoising process, and achieves superior denoising performance in both qualitative and quantitative evaluations while maintaining image details and outperforms existing state-of the-art methods.

Abstract

PET imaging is a powerful modality offering quantitative assessments of molecular and physiological processes. The necessity for PET denoising arises from the intrinsic high noise levels in PET imaging, which can significantly hinder the accurate interpretation and quantitative analysis of the scans. With advances in deep learning techniques, diffusion model-based PET denoising techniques have shown remarkable performance improvement. However, these models often face limitations when applied to volumetric data. Additionally, many existing diffusion models do not adequately consider the unique characteristics of PET imaging, such as its 3D volumetric nature, leading to the potential loss of anatomic consistency. Our Conditional Score-based Residual Diffusion (CSRD) model addresses these issues by incorporating a refined score function and 3D patch-wise training strategy, optimizing the model for efficient volumetric PET denoising. The CSRD model significantly lowers computational demands and expedites the denoising process. By effectively integrating volumetric data from PET and MRI scans, the CSRD model maintains spatial coherence and anatomical detail. Lastly, we demonstrate that the CSRD model achieves superior denoising performance in both qualitative and quantitative evaluations while maintaining image details and outperforms existing state-of-the-art methods.

Volumetric Conditional Score-based Residual Diffusion Model for PET/MR Denoising

TL;DR

The Conditional Score-based Residual Diffusion (CSRD) model is presented, which significantly lowers computational demands and expedites the denoising process, and achieves superior denoising performance in both qualitative and quantitative evaluations while maintaining image details and outperforms existing state-of the-art methods.

Abstract

PET imaging is a powerful modality offering quantitative assessments of molecular and physiological processes. The necessity for PET denoising arises from the intrinsic high noise levels in PET imaging, which can significantly hinder the accurate interpretation and quantitative analysis of the scans. With advances in deep learning techniques, diffusion model-based PET denoising techniques have shown remarkable performance improvement. However, these models often face limitations when applied to volumetric data. Additionally, many existing diffusion models do not adequately consider the unique characteristics of PET imaging, such as its 3D volumetric nature, leading to the potential loss of anatomic consistency. Our Conditional Score-based Residual Diffusion (CSRD) model addresses these issues by incorporating a refined score function and 3D patch-wise training strategy, optimizing the model for efficient volumetric PET denoising. The CSRD model significantly lowers computational demands and expedites the denoising process. By effectively integrating volumetric data from PET and MRI scans, the CSRD model maintains spatial coherence and anatomical detail. Lastly, we demonstrate that the CSRD model achieves superior denoising performance in both qualitative and quantitative evaluations while maintaining image details and outperforms existing state-of-the-art methods.
Paper Structure (9 sections, 7 equations, 3 figures, 1 table)

This paper contains 9 sections, 7 equations, 3 figures, 1 table.

Table of Contents

  1. Introduction
  2. Method
  3. Score-based Diffusion model for Residuals
  4. Patch-wise Training of Volumetric Conditional Score-based Diffusion Model
  5. Experiments
  6. Dataset Description
  7. Implementation Details
  8. Results
  9. Conclusion and Discussion

Figures (3)

  • Figure 1: A overview of volumetric PET denoising using a conditional score-based residual diffusion model. (A) The normal-dose PET volumes undergo Poisson thinning to simulate a low-dose PET scan scenario. Subsequently, the residual of normal-dose and low-dose PET volumes is generated. The residual, low-dose PET, and MRI volumes were split into smaller patches along with their respective spatial coordinates. (B) During the forward diffusion process, these patches undergo a noise addition process with time-dependent scheduling, represented by $\sigma(t)$. Then, the 3D U-net is trained for a score-matching function by removing an additive Gaussian noise with patch conditions of low-dose PET, MRI, and coordinates. (C) The trained network samples the residual of the entire volume from the Gaussian noise via the reverse diffusion process conditioned by entire low-dose PET and MRI associated with coordinates.
  • Figure 2: The representative PET images. The proposed conditional score-based diffusion model successfully improved the quality of PET volume while preserving fine anatomical details.
  • Figure 3: The images and error map in the unseen level of noise (1/10-dose)