서강대학교 전자공학과

Winner of the ‘Young Investigator Award’ at the 19th World Federation 

for Ultrasound in Medicine and Biology (WFUMB) Congress

▲ Department of Electronic Engineering, advised by Prof. Yangmo Yoo, Hyunwoo Cho

Hyunwoo Cho (Department of Electronic Engineering, advised by Prof. Yangmo Yoo), received the Young Investigator Award (YIA) at the 19th World Federation for Ultrasound in Medicine and Biology (WFUMB) Congress, held in Muscat, Oman, from November 4th to 7th.

At the conference, young researchers from medical ultrasound societies across each continent were nominated, and awards were determined through research presentations. Cho participated as a selected nominee by the Asian Federation of Societies for Ultrasound in Medicine and Biology (AFSUMB).

The title of Cho's research presentation was 'An Unsupervised Deep Beamformer for High-Quality Ultrafast Ultrasound Imaging.' This study proposed a novel method of unsupervised learning to dramatically improve the signal quality of plane wave ultrasound imaging.

Although plane wave ultrasound imaging enables exceptionally high frame rates and has been widely used for various clinical applications such as shear wave elastography and ultrafast perfusion imaging, it suffers from artifacts, noise, and low contrast caused by the characteristics of unfocused plane wave signals. Recent studies have highlighted deep learning-based image reconstruction as a promising solution to mitigate these limitations. However, most existing deep learning-based ultrasound image reconstruction techniques rely on supervised learning. Acquiring high-quality target data for supervised learning in the medical ultrasound field is extremely challenging, and the imperfect target data has been a significant hindrance to the performance of deep learning models.

This study proposed a new unsupervised learning method that can be trained with low-quality signals alone and showed significantly better performance in suppressing artifacts and noise and in improving contrast and resolution compared to existing supervised and self-supervised learning methods. The study also demonstrated that the deep learning model trained with this unsupervised learning method could be applied to advanced medical ultrasound applications like ultrafast perfusion imaging. This research is expected to be a foundational technology easily applicable to existing medical ultrasound imaging systems and to yield additional research outcomes in various medical ultrasound applications.

Cho expressed his gratitude: “I would like to thank my advisor Prof. Yangmo Yoo, as well as my lab seniors and colleagues who have supported and cheered me on. I will continue to work hard to conduct good research in the future.“

□ Organized by: World Federation for Ultrasound in Medicine and Biology (WFUMB)

□ Award: The Young Investigator Prize

□ Presentation Title: An Unsupervised Deep Beamformer for High-Quality Ultrafast Ultrasound Imaging