Using Artificial Neural Networks with multiple ECG leads (MLII and V1) achieved a 99.5% accuracy for ECG signal classification.
Does using multiple leads (MLII and V1) improve the accuracy of ECG signal classification with Artificial Neural Networks compared to using MLII alone?
Including a precordial lead (V1) alongside the main lead II (MLII) provides a 2D view of electrical patterns, allowing an Artificial Neural Network to achieve 99.5% accuracy in ECG signal classification.
This paper introduces the use of ECG signals from multiple leads to improve the accuracy of ECG signal classification with Artificial Neural Networks (ANN). The current methods commonly proposed rely on advanced signal processing or statistical analysis of the main lead II (MLII) in order to extract features that serve as a description of the signal. MLII, while being the most easily obtained ECG signal, does not contain a complete description of the electrical activity of the heart. Therefore, we propose to include a precordial lead, V1, from the Standard 12-Lead ECG system, to give the neural network a 2D view of the electrical patterns that arise during heart activation. This method was shown to be 99.5% accurate.
Perez et al. (Sun,) conducted a other in ECG signal classification. Artificial Neural Networks using multiple leads (MLII and V1) vs. Methods relying only on main lead II (MLII) was evaluated on ECG signal classification accuracy. Using Artificial Neural Networks with multiple ECG leads (MLII and V1) achieved a 99.5% accuracy for ECG signal classification.