A pneumonia detection system developed by a team from National Cheng Kung University in southern Taiwan’s Tainan City was honored in the health category of the online COVID-19 Global Hackathon.
Comprising four graduate students at NCKU’s Department of Computer Science and Information Engineering, MedCheX employed machine learning to design a model that can automatically detect the presence of pneumonia based on 1,400 positive and negative chest X-rays provided by the Department of Medical Imaging at NCKU Hospital.
According to Chiang Jung-hsien, a distinguished professor in the department, the system can analyze chest X-rays and send an electronic alert if it detects a suspected case of pneumonia, all without requiring a radiologist.
The scholar said the system has already been tested by doctors at NCKU Hospital, who found it achieved 92 percent accuracy with real patients. The team is working on incorporating CT scans or MRIs to further improve reliability, he added.
Sponsored by global corporate heavyweights including U.S.-headquartered Facebook Inc., Giphy Inc., Microsoft Corp. and Oracle Corp., the hackathon encouraged developers to build software solutions for managing the COVID-19 pandemic.
A total of 1,560 submissions were made March 26-30 in seven categories. Of these, 89 projects were honored, with MedCheX the only one from Taiwan. (SFC-E)