In this work, we apply HDR techniques to ultrasound imaging, where we combine ultrasound images acquired at different power levels to improve the level of detail visible in the final image. Our results strongly suggest that HDR-US imaging can improve the utility of ultrasound in image-based diagnosis and procedure guidance.
- High Dynamic Range Ultrasound Imaging, A. Degirmenci, D.P. Perrin, and R.D. Howe, Int J CARS (2018). https://link.springer.com/article/10.1007/s11548-018-1729-3 (Free, view-only version: http://rdcu.be/JfOg) Recipient of the Best Paper Award at IPCAI 2018
@Article{Degirmenci2018,
author="Degirmenci, Alperen
and Perrin, Douglas P.
and Howe, Robert D.",
title="High dynamic range ultrasound imaging",
journal="International Journal of Computer Assisted Radiology and Surgery",
year="2018",
month="May",
day="01",
volume="13",
number="5",
pages="721--729",
issn="1861-6429",
doi="10.1007/s11548-018-1729-3",
url="https://doi.org/10.1007/s11548-018-1729-3"
}
Download the repo to your machine:
git clone --recurse-submodules https://github.com/adegirmenci/HDR-US.git
From the root project directory, run the install script:
installHDRUS
- Image Processing Toolbox
- Parallel Computing Toolbox (optional)
This code was tested using MATLAB 2016b and 2017b.
We use F. Banterle's HDR Toolbox (https://github.com/banterle/HDR_Toolbox). The DebevecCRF function is modified to return two extra variables, logE and stack samples.
Run the MATLAB script runHDRUS.m:
runHDRUS
Select the dataset directory from the dialog.
To save results, set the flag
saveResults = true;
HDR-US was developed at the Harvard Biorobotics Lab. Licensed under the GNU General Public License Version 3 (GPLv3).