Korean government, private hospital researchers develop medical imaging technology using machine learning
The Korea Institute of Machinery and Materials, a research body under South Korea’s Ministry of Science and ICT, said its researchers have developed a cardiovascular event prediction model that enhances the speed and accuracy of disease diagnosis.
Researchers led by Dr Jong-won Park, head of the institute’s Department of Reliability Assessment, collaborated with the cardiology research group at Daejeon St. Mary’s Hospital.
KIMM said in a statement the research team integrated the big data deep learning technology used in checking the reliability of mechanical parts and equipment into ultrasound imaging equipment.
WHY IT MATTERS
According to KIMM, the new technology utilises a graphics processing unit to achieve a diagnosis time of 30 minutes with 80% accuracy.
Using AI deep learning, the researchers came up with the model through automated analysis of aortic atherosclerotic plaque. The research institute said they were “successful in confirming the effectiveness of such methods.”
In their study, the research team adopted a fresh approach toward creating a deep learning model that can be deployed to classify aortic plaque and measure plaque thickness. They applied standard machine learning techniques, such as autoencoder and U-Net models, to differentiate ultrasound images of the aortic wall, which was identified to confirm the conditions of any aortic atherosclerotic plaque – a risk factor for stroke.
KIMM noted that the researchers plan to modify the deep learning model to improve the accuracy of aortic plaque analysis. They also intend to expand the technology to be used along with imaging data for spotting faults and failures in building virtual engineering platforms for manufacturing future transportation equipment parts.
THE LARGER TREND
Researchers at the University of Western Australia are also trying to come up with an AI-based tool to better detect plaque in heart computed tomography scans. They envisioned the tool to check if a plaque has narrowed coronary arteries, identifying patients who are most at risk of adverse cardiovascular events.
A multi-disciplinary research team involving Australian listed firm InteliCare, Macquarie University and the University of Sydney is also working with machine learning technologies to create a tool that predicts risks of chronic disease and mental health deterioration among aged care patients.
Last year, researchers from Penn State University and Houston Methodist Hospital presented a machine learning-based tool that utilises a smartphone to quickly gauge facial movements and speech for signs of stroke.
ON THE RECORD
“Until now, users were required to have complex data analysis skills to determine the failure or lifespan of mechanical parts and equipment, but now they can easily access such information through various open sources,” Dr Park was quoted as saying.
“Imaging technology is expected to be implemented across various fields in the future, such as in the detection of various diseases and in the development of prediction models for the reliability of various parts and equipment,” he added.