Pseudo anomalies enhanced deep support vector data description for electrocardiogram quality assessment

Xunhua Huang, Fengbin Zhang, Haoyi Fan, Huihui Chang, Bing Zhou, Zuoyong Li

Abstract: Highlights•Self-supervised ECG quality assessment using anomaly detection ideas.•Multiple noise combinations to simulate real ECG noise.•Simulated noise as pseudo anomalies to learn better hypersphere boundary.•Using the generative relationship between ECG and noise to assist in model learning.