Cardiac abnormalities are a leading cause of death and their early diagnosis is important for providing timely interventions. The performance of modern, automated algorithms for classifying cardiac abnormalities from 12-lead ECG data was recently investigated at the 2020 PhysioNet / Computing in Cardiology Challenge. As part of the 2020 challenge the world’s largest open access database of 12-lead ECGs was released, which was drawn from three continents with diverse and distinctly different populations. In this talk, I will present our winning entry to the 2020 PhysioNet Challenge that combines both traditional hand-crafted ECG features, together with a novel deep learning architecture that learns discriminative feature representations.
Want to sponsor this event? contact us.