Jin Yong Jeon of Hanyang University discussed a technique using a machine learning algorithm to diagnose pneumonia through passive listening in his session, “Pneumonia diagnosis algorithm based on room impulse responses using cough sounds” during the 183rd Meeting of the Acoustical Society of America.
Pneumonia is one of the world’s leading causes of death and affects over a million people a year in the United States. The disease disproportionately impacts children, older adults, and hospitalized patients. To give them the greatest chance at recovery, it is crucial to catch and treat it early. Existing diagnosis methods consist of a range of blood tests and chest scans, and a doctor needs to suspect pneumonia before ordering them.
Jeon and fellow researchers developed a machine learning algorithm to identify cough sounds and determine whether the patient was suffering from pneumonia. Because every room and recording device is different, they augmented their recordings with room impulse responses, which measure how the acoustics of a space react to different sound frequencies. By combining this data with the recorded cough sounds, the algorithm can work in any environment.
“Automatically diagnosing a health condition through information on coughing sounds that occur continuously during daily life will facilitate non-face-to-face treatment,” says Jeon. “It will also be possible to reduce overall medical costs.”
Currently, one company has plans to apply this machine learning algorithm for remote patient monitoring. The team is also looking to implement it as an app for in-home care, and they plan to make the experience simpler and more user-friendly.
“Our research team is planning to automate each step-by-step process that is currently performed manually to improve convenience and applicability,” says Jeon.
Featured Image: A machine learning algorithm identifies cough sounds and determines whether the subject is suffering from pneumonia. Credit: Jin Yong Jeon