A Lightweight Deep Learning Model for Fast Electrocardiographic Beats Classification With a Wearable Cardiac Monitor: Development and Validation Study

Abstract

Background: Electrocardiographic (ECG) monitors have been widely used for diagnosing cardiac arrhythmias for decades. However, accurate analysis of ECG signals is difficult and time-consuming work because large amounts of beats need to be inspected. In order to enhance ECG beat classification, machine learning and deep learning methods have been studied. However, existing studies have limitations in model rigidity, model complexity, and inference speed.

Objective: To classify ECG beats effectively and efficiently, we propose a baseline model with recurrent neural networks (RNNs). Furthermore, we also propose a lightweight model with fused RNN for speeding up the prediction time on central processing units (CPUs).

Methods: We used 48 ECGs from the MIT-BIH (Massachusetts Institute of Technology-Beth Israel Hospital) Arrhythmia Database, and 76 ECGs were collected with S-Patch devices developed by Samsung SDS. We developed both baseline and lightweight models on the MXNet framework. We trained both models on graphics processing units and measured both models' inference times on CPUs.

Results: Our models achieved overall beat classification accuracies of 99.72% for the baseline model with RNN and 99.80% for the lightweight model with fused RNN. Moreover, our lightweight model reduced the inference time on CPUs without any loss of accuracy. The inference time for the lightweight model for 24-hour ECGs was 3 minutes, which is 5 times faster than the baseline model.

Conclusions: Both our baseline and lightweight models achieved cardiologist-level accuracies. Furthermore, our lightweight model is competitive on CPU-based wearable hardware.

Keywords: ECG classification; deep learning; fused recurrent neural network; path-type ECG sensor system; recurrent neural network.

Figure 1

Usage of S-Patch for Samsung SDS Cardio.

Figure 2

Data preprocessing of electrocardiograms: (A) full steps from downsampling to short-time Fourier transform; (B) an example of a 3-beat electrocardiographic segment for each class. ECG: electrocardiogram; RNN: recurrent neural network.

Figure 3

Examples of electrocardiographic signal noise: (A) loose contact, (B) motion artifact, (C) muscular activation interference, (D) baseline wandering, (E) alternating current (AC) interference (low signal-to-noise ratio), (F) AC interference (high signal-to-noise ratio).

Figure 4

Model architecture (general).

Figure 5

Normalized confusion matrix of the recurrent neural network (RNN) and fused RNN models.

Figure 6

Beats in rhythms with different shapes. SVT: supraventricular tachycardia; S: supraventricular ectopic; N: nonectopic.