Enhancing atrial fibrillation classification from single-lead electrocardiogram signals using attention-based networks and generative adversarial networks with density-based clustering

Godwin Msigwa, Ester Ntambala, Jaeseok Yun

Abstract: Highlights•MuDANet: Attention-based network for AFR and non-AFR rhythm distinction.•MuDANet uses dual CNN+Bi-LSTM and attention for better AFR classification.•DBSCAN-GAN augments data, optimizes GANs, and tackles outliers in training•DBSCAN-GAN generates quality ECG samples to enhance AFR classification.