Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Diffusion Reconstruction of Ultrasound Images with Informative Uncertainty

Yuxin Zhang, Clément Huneau, Jérôme Idier, Diana Mateus

TL;DR

This work adapts Denoising Diffusion Restoration Models (DDRM) to incorporate ultrasound physics through a linear direct model and an unsupervised fine-tuning of the prior diffusion model, demonstrating the efficacy of this approach in achieving high-quality image reconstructions from a single plane wave input and in comparison to state-of-the-art methods.

Abstract

Despite its wide use in medicine, ultrasound imaging faces several challenges related to its poor signal-to-noise ratio and several sources of noise and artefacts. Enhancing ultrasound image quality involves balancing concurrent factors like contrast, resolution, and speckle preservation. In recent years, there has been progress both in model-based and learning-based approaches to improve ultrasound image reconstruction. Bringing the best from both worlds, we propose a hybrid approach leveraging advances in diffusion models. To this end, we adapt Denoising Diffusion Restoration Models (DDRM) to incorporate ultrasound physics through a linear direct model and an unsupervised fine-tuning of the prior diffusion model. We conduct comprehensive experiments on simulated, in-vitro, and in-vivo data, demonstrating the efficacy of our approach in achieving high-quality image reconstructions from a single plane wave input and in comparison to state-of-the-art methods. Finally, given the stochastic nature of the method, we analyse in depth the statistical properties of single and multiple-sample reconstructions, experimentally show the informativeness of their variance, and provide an empirical model relating this behaviour to speckle noise. The code and data are available at: (upon acceptance).

Diffusion Reconstruction of Ultrasound Images with Informative Uncertainty

TL;DR

This work adapts Denoising Diffusion Restoration Models (DDRM) to incorporate ultrasound physics through a linear direct model and an unsupervised fine-tuning of the prior diffusion model, demonstrating the efficacy of this approach in achieving high-quality image reconstructions from a single plane wave input and in comparison to state-of-the-art methods.

Abstract

Despite its wide use in medicine, ultrasound imaging faces several challenges related to its poor signal-to-noise ratio and several sources of noise and artefacts. Enhancing ultrasound image quality involves balancing concurrent factors like contrast, resolution, and speckle preservation. In recent years, there has been progress both in model-based and learning-based approaches to improve ultrasound image reconstruction. Bringing the best from both worlds, we propose a hybrid approach leveraging advances in diffusion models. To this end, we adapt Denoising Diffusion Restoration Models (DDRM) to incorporate ultrasound physics through a linear direct model and an unsupervised fine-tuning of the prior diffusion model. We conduct comprehensive experiments on simulated, in-vitro, and in-vivo data, demonstrating the efficacy of our approach in achieving high-quality image reconstructions from a single plane wave input and in comparison to state-of-the-art methods. Finally, given the stochastic nature of the method, we analyse in depth the statistical properties of single and multiple-sample reconstructions, experimentally show the informativeness of their variance, and provide an empirical model relating this behaviour to speckle noise. The code and data are available at: (upon acceptance).
Paper Structure (21 sections, 14 equations, 8 figures, 1 table)

This paper contains 21 sections, 14 equations, 8 figures, 1 table.

Table of Contents

  1. Introduction
  2. METHODS
  3. Forward Ultrasound Imaging Model
  4. From Classical Beamforming to Inverse Problems
  5. Diffusion Restoration Models
  6. Diffusion Reconstruction of Ultrasound Images
  7. Experimental Setup
  8. Training Datasets
  9. Validation Datasets
  10. Inverse Problem Model Parameters
  11. Evaluation Metrics
  12. Experimental Validation and Results
  13. Phantom based performance assesment
  14. Sensitivity of Deno and DRUS to the iteration step count for single-sample reconstructions
  15. Statistical behavior of the reconstruction
  16. ...and 6 more sections

Figures (8)

  • Figure 1: Sensitivity of Deno and DRUS to the iteration step count in the sampling process, evaluated on PICMUS, and compared to DAS.
  • Figure 2: Statistical behaviour of Deno and DRUS compared with the DAS1 image on PICMUS.
  • Figure 3: Quantitative comparison of single- and multiple-sample images on PICMUS. In the multiple-sample case, each sample relies on 50 it.
  • Figure 4: Qualitative comparison of the reconstructed phantom-based PICMUS images. All images are in decibels with a dynamic range [-60,0].
  • Figure 5: Comparison of reconstructed images on the PICMUS in vivo datasets. All images are in decibels with a dynamic range [-60,0].
  • ...and 3 more figures