Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Baseline Drift Tolerant Signal Encoding for ECG Classification with Deep Learning

Robert O Shea, Prabodh Katti, Bipin Rajendran

TL;DR

The paper addresses the vulnerability of ECG deep learning classification to artefacts such as baseline drift, rescaling, and noise. It introduces Derived Peak (DP) encoding, a parameter-free, derivative-based, event-driven representation that emits signed spikes at zero-crossings of the signal's first and/or second derivatives, invariant to shifts and gains. When applied to the PTB-XL dataset with a 1D-ResNet-18 classifier, DP maintains high discriminative performance under artefacts and outperforms several discretisation- and TEM-based encodings in robustness, achieving AUCs around the 0.88–0.91 range under challenging perturbations while keeping spike rates low. These results suggest DP as a practical, energy-efficient encoding for robust, edge-ready ECG analysis and can be combined with other schemes to further enhance artefact resilience.

Abstract

Common artefacts such as baseline drift, rescaling, and noise critically limit the performance of machine learningbased automated ECG analysis and interpretation. This study proposes Derived Peak (DP) encoding, a non-parametric method that generates signed spikes corresponding to zero crossings of the signals first and second-order time derivatives. Notably, DP encoding is invariant to shift and scaling artefacts, and its implementation is further simplified by the absence of userdefined parameters. DP encoding was used to encode the 12-lead ECG data from the PTB-XL dataset (n=18,869 participants) and was fed to 1D-ResNet-18 models trained to identify myocardial infarction, conductive deficits and ST-segment abnormalities. Robustness to artefacts was assessed by corrupting ECG data with sinusoidal baseline drift, shift, rescaling and noise, before encoding. The addition of these artefacts resulted in a significant drop in accuracy for seven other methods from prior art, while DP encoding maintained a baseline AUC of 0.88 under drift, shift and rescaling. DP achieved superior performance to unencoded inputs in the presence of shift (AUC under 1mV shift: 0.91 vs 0.62), and rescaling artefacts (AUC 0.91 vs 0.79). Thus, DP encoding is a simple method by which robustness to common ECG artefacts may be improved for automated ECG analysis and interpretation.

Baseline Drift Tolerant Signal Encoding for ECG Classification with Deep Learning

TL;DR

The paper addresses the vulnerability of ECG deep learning classification to artefacts such as baseline drift, rescaling, and noise. It introduces Derived Peak (DP) encoding, a parameter-free, derivative-based, event-driven representation that emits signed spikes at zero-crossings of the signal's first and/or second derivatives, invariant to shifts and gains. When applied to the PTB-XL dataset with a 1D-ResNet-18 classifier, DP maintains high discriminative performance under artefacts and outperforms several discretisation- and TEM-based encodings in robustness, achieving AUCs around the 0.88–0.91 range under challenging perturbations while keeping spike rates low. These results suggest DP as a practical, energy-efficient encoding for robust, edge-ready ECG analysis and can be combined with other schemes to further enhance artefact resilience.

Abstract

Common artefacts such as baseline drift, rescaling, and noise critically limit the performance of machine learningbased automated ECG analysis and interpretation. This study proposes Derived Peak (DP) encoding, a non-parametric method that generates signed spikes corresponding to zero crossings of the signals first and second-order time derivatives. Notably, DP encoding is invariant to shift and scaling artefacts, and its implementation is further simplified by the absence of userdefined parameters. DP encoding was used to encode the 12-lead ECG data from the PTB-XL dataset (n=18,869 participants) and was fed to 1D-ResNet-18 models trained to identify myocardial infarction, conductive deficits and ST-segment abnormalities. Robustness to artefacts was assessed by corrupting ECG data with sinusoidal baseline drift, shift, rescaling and noise, before encoding. The addition of these artefacts resulted in a significant drop in accuracy for seven other methods from prior art, while DP encoding maintained a baseline AUC of 0.88 under drift, shift and rescaling. DP achieved superior performance to unencoded inputs in the presence of shift (AUC under 1mV shift: 0.91 vs 0.62), and rescaling artefacts (AUC 0.91 vs 0.79). Thus, DP encoding is a simple method by which robustness to common ECG artefacts may be improved for automated ECG analysis and interpretation.
Paper Structure (7 sections, 1 equation, 3 figures, 2 tables)

This paper contains 7 sections, 1 equation, 3 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Derived Peak Encoding
  4. Experiments and Data
  5. Results
  6. Discussion
  7. Conclusion

Figures (3)

  • Figure 1: DP encoding scheme, for first and second-order derivatives, for a sample ECG.
  • Figure 2: Experimental flow to test ResNet classification performance of spike-encoded ECG signals. To model the continuous low-latency ECG monitoring setting, subseries of 2,560ms duration were extracted from random timepoints. Signal pre-processing was limited to 3-point moving average (MA) for reduction of high-frequency noise.
  • Figure 3: ECG sample corrupted with artefacts used in the study