The Irish Centre for High-Performance Computing (ICHEC) has approved a third major COVID-19 research project on the national high-performance computer (HPC), known as Kay.
Led by Dr. Aaron Golden from the National University of Ireland, Galway, the team plans to develop new artificial intelligence (AI) software to expedite the diagnosis of COVID-19 from chest CT scans.
Nasal and throat swabs, the traditional methods for detecting COVID-19, are not 100% accurate, and often patients don't receive their results for a day or two. However, using a CT scan, a radiologist can tell in less than an hour if a patient has lung lesions indicative of COVID-19, Golden said.
The question becomes the following: Are the lesions due to pneumonia, a pulmonary disorder, or another lung condition? That's where AI comes in, as well as the necessity of Kay's supercomputing power. A convolutional neural network classifies CT scans into likely COVID-19 or non-COVID-19 groups, and then a deep-learning algorithm standardizes the thousands of training CT scans.
Golden's project was funded by the Health Research Board as part of the COVID-19 Pandemic Rapid Response Funding Call.
