Enhanced ECG Signal Classification Using Multi‐Branch Convolutions and Mamba Blocks With State‐Space Models

ABSTRACT Electrocardiogram (ECG) classification is crucial for the timely detection of cardiac arrhythmias and related heart conditions. This study presents an innovative deep‐learning model that integrates convolutional neural networks (CNNs), long short‐term memory (LSTM) units, and Mamba blocks with state‐space models (SSMs) to enhance the precision of ECG classification. This model addresses the limitations of transformer‐based approaches, particularly their high computational complexity and inefficiencies during the inference phase, exacerbated by their self‐attention mechanisms when processing long sequences. Evaluations on the MIT‐BIH dataset reveal that our model achieves outstanding classification accuracy, with a computational efficiency of 99.26%. Integrating the multibranch convolutional approach with Mamba blocks provides a robust and efficient framework for ECG signal analysis, offering a promising solution for real‐time cardiovascular monitoring and diagnosis.