This paper presents an AI system for rapid COVID-19 detection using chest CT scans. The system, a deep convolutional neural network, was trained and evaluated on a large dataset of over 10,000 CT volumes from various pneumonia types and healthy controls. The system achieved high accuracy (AUC of 97.81% on the test cohort), outperforming five radiologists in reader studies, while being significantly faster. The study also compared CT and X-ray performance and included an interpretation of the network's decision-making process.
