SoC-Based Implementation of 1D Convolutional Neural Network for 3-Channel ECG Arrhythmia Classification via HLS4ML

Real-time monitoring of one-dimensional (1D) biopotentials, such as electrocardiograms (ECG), necessitates effective feature extraction and classification, a strength of Deep Learning (DL) algorithms. Designing 1D Convolutional Neural Network (1D CNN) accelerators for biopotential classification via open-source codesign workflows, particularly High-Level Synthesis for Machine Learning (HLS4ML), offers advantages over GPU-based or Cloud-based solutions, including high performance, low latency, low power consumption, swift time-to-market, and cost-effectiveness. We present an implementation of a quantized-pruned (QP) 1D CNN model on the PYNQ Z2 SoC using HLS4ML by seamlessly deploying its soft IP core generated via Vivado Accelerator backend, showcasing the efficacy of Quantization Aware Training (QAT) in reducing power consumption to 1.655W from 1.823W. Our approach demonstrates improved area consumption, resource utilization, and inferences per second compared to the baseline (B) 1D CNN model, with a controlled 4% or less reduction in weighted Accuracy, Precision, Recall, and F1-score, revealing the nuanced trade-offs between performance metrics and system efficiency for real-time 3-Channel ECG Arrhythmia classification.