Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

When Diffusion MRI Meets Diffusion Model: A Novel Deep Generative Model for Diffusion MRI Generation

Xi Zhu, Wei Zhang, Yijie Li, Lauren J. O'Donnell, Fan Zhang

TL;DR

This work tackles the challenge of obtaining high-quality diffusion MRI data by proposing a novel Latent Diffusion Model–based framework that generates 7T-like dMRI from 3T data. It combines rotation-invariant spherical harmonics (RISH) with a transfer-learning VQ-VAE and a classifier-free guided diffusion process in latent space, augmented by a super-resolution module to align spatial resolutions. The approach demonstrates superior NMSE and SSIM metrics over CNN- and GAN-based baselines on HCP data, with ablation studies highlighting the importance of fine-tuning and the SR component. The method holds potential to reduce scanner time and cost while elevating imaging standards for diffusion MRI in clinical and research settings.

Abstract

Diffusion MRI (dMRI) is an advanced imaging technique characterizing tissue microstructure and white matter structural connectivity of the human brain. The demand for high-quality dMRI data is growing, driven by the need for better resolution and improved tissue contrast. However, acquiring high-quality dMRI data is expensive and time-consuming. In this context, deep generative modeling emerges as a promising solution to enhance image quality while minimizing acquisition costs and scanning time. In this study, we propose a novel generative approach to perform dMRI generation using deep diffusion models. It can generate high dimension (4D) and high resolution data preserving the gradients information and brain structure. We demonstrated our method through an image mapping task aimed at enhancing the quality of dMRI images from 3T to 7T. Our approach demonstrates highly enhanced performance in generating dMRI images when compared to the current state-of-the-art (SOTA) methods. This achievement underscores a substantial progression in enhancing dMRI quality, highlighting the potential of our novel generative approach to revolutionize dMRI imaging standards.

When Diffusion MRI Meets Diffusion Model: A Novel Deep Generative Model for Diffusion MRI Generation

TL;DR

This work tackles the challenge of obtaining high-quality diffusion MRI data by proposing a novel Latent Diffusion Model–based framework that generates 7T-like dMRI from 3T data. It combines rotation-invariant spherical harmonics (RISH) with a transfer-learning VQ-VAE and a classifier-free guided diffusion process in latent space, augmented by a super-resolution module to align spatial resolutions. The approach demonstrates superior NMSE and SSIM metrics over CNN- and GAN-based baselines on HCP data, with ablation studies highlighting the importance of fine-tuning and the SR component. The method holds potential to reduce scanner time and cost while elevating imaging standards for diffusion MRI in clinical and research settings.

Abstract

Diffusion MRI (dMRI) is an advanced imaging technique characterizing tissue microstructure and white matter structural connectivity of the human brain. The demand for high-quality dMRI data is growing, driven by the need for better resolution and improved tissue contrast. However, acquiring high-quality dMRI data is expensive and time-consuming. In this context, deep generative modeling emerges as a promising solution to enhance image quality while minimizing acquisition costs and scanning time. In this study, we propose a novel generative approach to perform dMRI generation using deep diffusion models. It can generate high dimension (4D) and high resolution data preserving the gradients information and brain structure. We demonstrated our method through an image mapping task aimed at enhancing the quality of dMRI images from 3T to 7T. Our approach demonstrates highly enhanced performance in generating dMRI images when compared to the current state-of-the-art (SOTA) methods. This achievement underscores a substantial progression in enhancing dMRI quality, highlighting the potential of our novel generative approach to revolutionize dMRI imaging standards.
Paper Structure (12 sections, 10 equations, 3 figures, 2 tables)

This paper contains 12 sections, 10 equations, 3 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Method
  3. dMRI Datasets
  4. dMRI signal representation and reconstruction using RISH
  5. Latent diffusion model
  6. Experimental Comparisons
  7. Results
  8. SOTA comparison.
  9. Ablation experiment.
  10. Conclusion
  11. Acknowledgments.
  12. Disclosure of Interests.

Figures (3)

  • Figure 1: Overview of the proposed method.
  • Figure 2: Results for the RISH features and FA generated by different methods.
  • Figure 3: Difference maps of FA images.