SoC-Based Implementation of 1D Convolutional Neural Network for 3-Channel ECG Arrhythmia Classification via HLS4ML
Published in IEEE Embedded Systems Letters, 2024
This paper presents a SoC-based implementation of a 1-D convolutional neural network (1-D CNN) for 3-channel ECG arrhythmia classification using HLS4ML. It demonstrates the benefits of quantization-aware training (QAT) and high-level synthesis (HLS) in reducing power consumption while maintaining competitive performance metrics, offering an efficient, low-latency, and cost-effective solution for real-time ECG monitoring.
Recommended citation: Ahmad, F., & Zafar, S. (2024). "SoC-Based Implementation of 1-D Convolutional Neural Network for 3-Channel ECG Arrhythmia Classification via HLS4ML." IEEE Embedded Systems Letters, 16(4), 429-432. doi: 10.1109/LES.2024.3354081 https://ieeexplore.ieee.org/abstract/document/10399904