Skip to content
Open access

Inter-Patient Multi-Label Localization of Myocardial Infarction Territories Using 12-Lead ECG Signals

Aug 2026 · Engineering, Technology & Applied Science Research · 0 citations · 21 references

Abstract

The localization of Myocardial Infarction (MI) from Electrocardiogram (ECG) signals remains challenging due to inter-patient variability, class imbalance, and limited interpretability of current methods. This study presents a one-dimensional convolutional network, MobileECGNet, for multi-label, territory-wise MI localization using 12-lead ECG signals. The architecture incorporates depthwise separable convolutions, squeeze-and-excitation blocks, and temporal attention to capture inter-lead dependencies and temporal waveform characteristics. Experiments were conducted on the PTB-XL dataset using an inter-patient evaluation protocol with stratified splitting to preserve label distribution. ECG signals were standardized to fixed-length segments to ensure consistent input representation. The proposed method achieved a macro-AUC of 0.93, indicating strong localization performance. In addition, to analyze lead-wise contributions, a SHAP-based explainable AI technique was used, which exhibits alignment with clinically relevant ECG leads. Overall, the results demonstrate that the proposed framework provides an effective and interpretable solution for multi-label MI localization problems.

Read PDF

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.