FADE: Forecasting for anomaly detection on ECG

Paula Ruiz-Barroso, Francisco M. Castro, José Miranda, Denisa-Andreea Constantinescu, David Atienza, Nicolás Guil

Abstract: Highlights•New self-supervised deep learning model for ECG forecasting with a novel loss.•Domain adaptation with a new distance metric to detect ECG anomalies.•Thorough empirical study to validate the performance of the proposed approach.•Understandability study to unravel the knowledge acquired by the model.•Accurate anomaly or normal signal detection without seeing anomalies during training.

External IDs:dblp:journals/cmpb/RuizBarrosoCMCAG25