Home | Publications | DBF+26

UltraBend: Refraction-Aware Propagation Modeling for Speed of Sound Imaging in Ultrasound

MCML Authors

Abstract

The local speed of sound (SoS) of human soft tissues is a promising quantitative biomarker for the diagnosis of diseases such as breast cancer or fatty liver disease. One approach to estimate local SoS from pulse-echo ultrasound signals is the autofocusing method, which optimizes for the propagation model that best aligns beamformed signals recorded from different transmit--receive apertures. However, most pulse-echo autofocusing methods still rely on a straight-ray hypothesis, thus neglecting refraction effects occurring in heterogeneous media. To address this shortcoming, we introduce UltraBend, which replaces straight-line travel-time approximations with a refraction-aware bent-ray model for local SoS optimization. We compare our method with a state-of-the-art autofocusing method assuming straight-ray wave propagation and show that UltraBend reduces the mean absolute reconstruction SoS error relative to a straight-ray baseline on all evaluated phantoms, including a reduction from 8.51 to 3.70 m/s on an inclusion phantom. These results suggest that improved modeling of wave propagation can substantially improve pulse-echo SoS imaging.

inproceedings DBF+26

ASMUS @MICCAI 2026

7th International Workshop on Advances in Simplifying Medical UltraSound at the 29th International Conference on Medical Image Computing and Computer Assisted Intervention. Strasbourg, France, Sep 27-Oct 01, 2026. To be published. Preprint available.

Authors

F. Dülmer • S. Beuret • B. N. Frey • M. F. Azampour • N. Navab • J. Dahl

Links

URL

Research Area

 C1 | Medicine

BibTeXKey: DBF+26

#p-navab
Back to Top