EdgeECG: a lightweight edge-oriented network with dual criterion pruning for real-time ECG arrhythmia classification
Clinical Snapshot
PICO Framework
| P — Population | ECG signals from MIT-BIH arrhythmia dataset representing patients with various cardiac arrhythmias |
| I — Intervention | EdgeECG neural network with dual criterion pruning (DCP) strategy deployed on STM32F103 microcontroller |
| C — Comparator | Several representative ECG classification methods and baseline neural networks |
| O — Outcomes | Classification accuracy, model size (parameters), inference latency, energy consumption, and deployment feasibility on resource-constrained devices |
Bottom Line
EdgeECG represents a significant advancement in point-of-care ECG analysis, achieving 98.07% classification accuracy while requiring only 2680 parameters and minimal computational resources. The successful deployment on a basic microcontroller with 0.025-second inference time demonstrates practical viability for wearable cardiac monitoring. The 80.73% patient-independent accuracy, while lower, suggests reasonable generalizability across different patients. For Australian practice, this technology could enhance remote monitoring capabilities, particularly valuable for rural healthcare delivery and chronic disease management. However, clinical validation beyond benchmark datasets is needed before widespread adoption. The ultra-low power consumption (0.156mJ per inference) makes it suitable for battery-powered wearable devices, potentially improving patient compliance with continuous monitoring. Clinicians should note that while promising, this remains a research tool requiring further validation against clinical gold standards and regulatory approval before clinical implementation.
Key Findings
P Value: Not reported
Effect Size: 98.07% overall accuracy, 80.73% patient-independent accuracy
Primary Outcome: ECG arrhythmia classification accuracy
Nnt Or Sensitivity: 2680 parameters, 0.025s inference latency, 0.156mJ energy consumption
Confidence Interval: Not reported
Clinical Application
High feasibility for deployment on resource-constrained devices, demonstrated successful implementation on STM32F103 microcontroller Highly relevant for Australian telehealth initiatives and remote patient monitoring programs, particularly for rural and remote communities. Could support PBS-funded cardiac monitoring devices and align with RACGP guidelines for cardiovascular risk management Patients requiring continuous ECG monitoring, particularly those using wearable devices or remote monitoring systems
Abstract
Objective.Miniature electrocardiogram (ECG) devices can rapidly and accurately acquire real-time cardiac signals, enabling timely warnings for patients with heart disease. To achieve accurate arrhythmia classification on resource-constrained ECG edge devices, we propose EdgeECG, an ultra-lightweight neural network designed for deployment on low-power microcontrollers.Approach.EdgeECG is first designed with a compact convolutional architecture to ensure compatibility with resource-limited embedded platforms. A dual criterion pruning (DCP) strategy is then introduced to evaluate weight importance using both magnitude and median deviation for more precise model compression. In addition, quantization is applied to reduce storage cost and improve deployment efficiency on an STM32F103 microcontroller.Main results.Experimental results on the benchmark MIT-BIH arrhythmia dataset show that EdgeECG achieves an overall classification accuracy of 98.07%, outperforming several representative methods. In patient-independent record-level evaluation, the accuracy reaches 80.73%. The model contains only 2680 parameters and is successfully deployed on an STM32F103 microcontroller with 64 KB SRAM, achieving an inference latency of 0.025 s and an energy consumption of 0.156 mJ per inference. In addition, DCP reduces the number of non-zero parameters by nearly 50% while maintaining promising classification performance.Significance.EdgeECG provides an effective solution for five-class ECG arrhythmia classification on severely resource-constrained edge devices. Its compact architecture, effective pruning strategy, and successful deployment on an STM32F103 demonstrate its potential for low-power edge-based ECG analysis and wearable cardiac monitoring applications. The source code of EdgeECG is publicly available at:https://github.com/jyanhuang/EdgeECG-code.
References
- 1.Huang, J., Wang, C., Grau, A., Xiao, Y., Luo, S., Che, Y., & Li, Z. (2024). EdgeECG: a lightweight edge-oriented network with dual criterion pruning for real-time ECG arrhythmia classification. Physiological Measurement. https://doi.org/10.1088/1361-6579/ae696a
Related Research
Bioelectrochemistry (Amsterdam, Netherlands)
From data to diagnosis: A comprehensive review of machine learning-driven wearable sensors in healthcare.
2 Aug 2026
Ultrasound in medicine & biology
A Systematic Review: The Application of Attention Mechanisms in Medical Ultrasound Image Processing
2 Aug 2026
Journal of the American Medical Informatics Association : JAMIA
Explainable machine learning in healthcare: methods, interpretation, and applications for clinical research
2 Aug 2026
This content is for educational purposes for healthcare professionals only and does not constitute clinical advice. Clinical decisions should be based on individual patient assessment, current guidelines, and appropriate specialist consultation. Editorial Standards · Privacy Policy · Terms of Service