Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

UltraGS: Real-Time Physically-Decoupled Gaussian Splatting for Ultrasound Novel View Synthesis

Yuezhe Yang, Qingqing Ruan, Wenjie Cai, Yudang Dong, Dexin Yang, Xingbo Dong, Zhe Jin, Yong Dai

TL;DR

This work tackles the challenge of real-time novel-view synthesis for sensorless ultrasound by adapting Gaussian Splatting to acoustic imaging. It introduces depth-aware Gaussian disks with Dynamic Aperture Rectification and a PD Rendering operator that decouples geometry from light-like wave interactions using low-order spherical harmonics and first-order acoustic effects. The method yields state-of-the-art PSNR and SSIM while achieving real-time performance (~64.7 fps) across wild, phantom, and clinical datasets, demonstrating robustness to unconstrained probe motion. The approach offers a hardware-free, efficient pathway toward clinically useful 3D ultrasound visualization and navigation.

Abstract

Ultrasound imaging is a cornerstone of non-invasive clinical diagnostics, yet its limited field of view poses challenges for novel view synthesis. We present UltraGS, a real-time framework that adapts Gaussian Splatting to sensorless ultrasound imaging by integrating explicit radiance fields with lightweight, physics-inspired acoustic modeling. UltraGS employs depth-aware Gaussian primitives with learnable fields of view to improve geometric consistency under unconstrained probe motion, and introduces PD Rendering, a differentiable acoustic operator that combines low-order spherical harmonics with first-order wave effects for efficient intensity synthesis. We further present a clinical ultrasound dataset acquired under real-world scanning protocols. Extensive evaluations across three datasets demonstrate that UltraGS establishes a new performance-efficiency frontier, achieving state-of-the-art results in PSNR (up to 29.55) and SSIM (up to 0.89) while achieving real-time synthesis at 64.69 fps on a single GPU. The code and dataset are open-sourced at: https://github.com/Bean-Young/UltraGS.

UltraGS: Real-Time Physically-Decoupled Gaussian Splatting for Ultrasound Novel View Synthesis

TL;DR

This work tackles the challenge of real-time novel-view synthesis for sensorless ultrasound by adapting Gaussian Splatting to acoustic imaging. It introduces depth-aware Gaussian disks with Dynamic Aperture Rectification and a PD Rendering operator that decouples geometry from light-like wave interactions using low-order spherical harmonics and first-order acoustic effects. The method yields state-of-the-art PSNR and SSIM while achieving real-time performance (~64.7 fps) across wild, phantom, and clinical datasets, demonstrating robustness to unconstrained probe motion. The approach offers a hardware-free, efficient pathway toward clinically useful 3D ultrasound visualization and navigation.

Abstract

Ultrasound imaging is a cornerstone of non-invasive clinical diagnostics, yet its limited field of view poses challenges for novel view synthesis. We present UltraGS, a real-time framework that adapts Gaussian Splatting to sensorless ultrasound imaging by integrating explicit radiance fields with lightweight, physics-inspired acoustic modeling. UltraGS employs depth-aware Gaussian primitives with learnable fields of view to improve geometric consistency under unconstrained probe motion, and introduces PD Rendering, a differentiable acoustic operator that combines low-order spherical harmonics with first-order wave effects for efficient intensity synthesis. We further present a clinical ultrasound dataset acquired under real-world scanning protocols. Extensive evaluations across three datasets demonstrate that UltraGS establishes a new performance-efficiency frontier, achieving state-of-the-art results in PSNR (up to 29.55) and SSIM (up to 0.89) while achieving real-time synthesis at 64.69 fps on a single GPU. The code and dataset are open-sourced at: https://github.com/Bean-Young/UltraGS.

Paper Structure

This paper contains 21 sections, 12 equations, 11 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Methods
  4. Preliminaries of Gaussian Splatting
  5. Metric-Consistent Depth Recovery and Dynamic Aperture Rectification
  6. PD Rendering: A Differentiable Neural Acoustic Operator
  7. Experiments
  8. Datasets
  9. Experimental Setup
  10. Quantitative and Qualitative Evaluation
  11. Ablation Study and Robustness Analysis
  12. Conclusion
  13. Mathematical Proxy
  14. Justification of the Virtual Pinhole Proxy
  15. COLMAP for Ultrasound: Rationale and Pre-processing
  16. ...and 6 more sections

Figures (11)

  • Figure 1: Pipeline of UltraGS. (a) Gaussian primitives are assigned learnable fields of view for dynamic aperture rectification, enabling metric-consistent depth recovery. (Illustrated here is Case 1 from the Clinical Dataset.) (b) PD Rendering synthesizes new images by integrating low-order spherical harmonics with an efficient approximation of acoustic wave propagation (attenuation, reflection, and scattering).
  • Figure 2: Comparison of primitive representations. (a) Standard 3DGS suffers from view inconsistency and (b) significant depth offsets in acoustic scenes. (c) UltraGS utilizes 2D Gaussian disks to facilitate metric-consistent projection and superior structural fidelity.
  • Figure 3: A visual reconstruction comparison on Wild Dataset and Clinical Dataset.
  • Figure 4: Comparison of kidney sample visualization results from the Clinical Dataset
  • Figure 5: A visual reconstruction comparison on Ablation Study and Robustness Analysis.
  • ...and 6 more figures