Ming-Chun Huang, associate professor of data and computational science at Duke Kunshan University (left) with Wenyao Xu, professor of computer science and engineering at SUNY Buffalo (right)
By John Butcher
Professors from Duke Kunshan and SUNY Buffalo have led a team of students to success in an international Covid-19 healthcare technology competition.
They showed the feasibility of building an app that could use artificial intelligence to diagnose Covid-19 by detecting respiratory system impairments through sound to win the runner-up award in the IEEE Health Summit Covid-19 Data Hackathon.
“Our exploration could boost the transition from the existing expensive molecular testing methods to non-contact, effortless and fast testing schemes that would allow daily self-checking for Covid-19 using a smartphone,” said Ming-Chun Huang, associate professor of data and computational science at Duke Kunshan.
The team was led by Huang and Wenyao Xu, professor of computer science and engineering at SUNY Buffalo, with students from SUNY Buffalo, CU Denver and Case Western Reserve University. Together they developed an artificial intelligence model to analyze the characteristic sounds of Covid-19 within a cough, which came second in the competition to a project by the National Technical University of Athens. The model is a first step towards potentially building a smartphone app that would be able to listen to a cough or a spoken sentence and provide a diagnosis within seconds.
The potential of the team’s research is not limited to Covid-19, as diagnosis through sound could be applied to many other heath conditions, according to Huang.
“Voice is an important biomarker that can demonstrate a user’s health,” he said. “According to the voice production theory, in addition to the respiratory and articulatory organs, voice is also controlled by the motor function and high-level cognitive function in the human brain. Therefore, it is feasible to leverage the voice to detect some neural diseases, such as Parkinson’s Disease, and some mental health issues, such as depression and anxiety.”
Students were invited onto the project based on an interest in artificial intelligence and data science. During the research they furthered their understanding of big data processing, deep learning, and deep learning model visualization skills.
With funding from the Kunshan Government Research Fund to advance studies into digital healthcare, Huang plans to conduct further studies and commercialize the results of his research in Kunshan.
“In the next step, we plan to consider demographic factors such as age and gender, as well as health histories, to develop a more accurate and personalized Covid-19 detection model,” he said.
