ML-Net: Multi-Channel Lightweight Network for Detecting Myocardial Infarction

Area of Science:

• Biomedical Engineering
• Artificial Intelligence in Medicine

Background:

• Myocardial infarction (MI) detection is challenging due to complex electrocardiogram (ECG) waveforms.
• Existing automatic diagnosis models for MI often require significant computational resources, limiting their use on portable devices.

Purpose of the Study:

• To develop a lightweight model for real-time, portable MI detection.
• To address the limitations of current MI diagnostic tools in resource-constrained environments.

Main Methods:

• Proposed a novel multi-channel lightweight network (ML-Net) for MI detection.
• Utilized independent ECG leads as separate channels to preserve data integrity.
• Employed convolution kernels of heterogeneous sizes for efficient classification.

Main Results:

• ML-Net demonstrated superior performance in diagnosing MI compared to existing methods.
• The model requires significantly lower computational cost and memory footprint.
• Achieved accurate classification using limited lead data.

Conclusions:

• ML-Net offers a viable solution for portable, real-time MI detection.
• The model's efficiency enhances the applicability of IoMT devices for cardiac diagnostics.
• Facilitates wider adoption of advanced MI detection in resource-limited settings.