Journal
2022 IEEE INTERNATIONAL ULTRASONICS SYMPOSIUM (IEEE IUS)
Volume -, Issue -, Pages -Publisher
IEEE
DOI: 10.1109/IUS54386.2022.9957817
Keywords
elastography; shear wave speed; skeletal muscle; masking; Radon Transform
Categories
Funding
- NIH [T32GM007171, R01HD096361, R01EB022106, R37HL096023, R01EB033064]
- Duke University MEDx Pilot Project
Ask authors/readers for more resources
The Radon Transform (RT) is a commonly used method for estimating shear wave trajectory and speed in ultrasound shear wave elasticity imaging (SWEI). Researchers have also developed a deep neural network to automatically generate masks, reducing the need for manual mask drawing.
The Radon Transform (RT) approach is a common method used to estimate shear wave trajectory and speed in ultrasound shear wave elasticity imaging (SWEI). The RT calculates the sum of 2D spatiotemporal data amplitude under each potential linear trajectory to determine the trajectory with the greatest value. We divide the RT of data by the RT of a ones matrix to normalize by trajectory path length, enabling the use of arbitrary data masks. We demonstrate that masking can isolate the two simultaneous SH and SV shear wave modes observed in in vivo skeletal muscle data. 38 rotational SWEI acquisitions were collected in vastus lateralis muscle, for a total of 2736 space-time plots, and shear waves were identified by manually drawing masks. Using these labeled data, we trained a deep neural network to generate masks from a space-time plot to reduce the need to hand-draw masks in the future. On a held-out test case, 91% of predicted trajectories corresponded to a labeled shear wave, and estimated speeds had a mean absolute error of 7.6%. Despite frequent use of the RT method in the literature, no openly available code exists. We have released our code at https://github.com/fqjin/radon-transform.
Authors
I am an author on this paper
Click your name to claim this paper and add it to your profile.
Reviews
Recommended
No Data Available