On the redundancy of denoising in electrocardiogram reconstruction - Report - MDSpire

On the redundancy of denoising in electrocardiogram reconstruction

  • By

  • Ekenedirichukwu N. Obianom

  • Abdulhamed M. Jasim

  • Abubakar Sadiq Muhammad

  • G. André Ng

  • Xin Li

  • June 26, 2026

  • 0 min

Share

Clinical Report: The Necessity of Denoising in Electrocardiogram Reconstruction

Overview

This study evaluates the performance of a single-model pipeline (SMP) for ECG reconstruction compared to a multiple-model pipeline (MMP). Results indicate that the SMP achieves comparable performance.

Background

Electrocardiogram (ECG) reconstruction is crucial for synthesizing missing leads from recorded ECG data, which is essential for accurate cardiac diagnosis. Traditional methods often involve denoising steps that can distort signal morphology. This study explores whether a unified model can effectively reconstruct ECG signals without the need for denoising.

Data Highlights

Pipeline TypeAverage CorrelationMemory Requirement
Single-Model Pipeline (SMP)Above 0.85Lower
Multiple-Model Pipeline (MMP)Above 0.85Over five times higher

Key Findings

  • The SMP performs comparably to the MMP with average correlations above 0.85.
  • Performance of the SMP remains stable across varying noise levels.
  • Transfer learning can enhance SMP performance on abnormal ECGs with minimal data.
  • The MMP requires significantly more memory due to the use of multiple models.
  • Eliminating denoising steps can streamline the reconstruction process.

Clinical Implications

The findings suggest that a single-model approach to ECG reconstruction can reduce computational demands.

Conclusion

The study supports the viability of using a single-model pipeline for ECG reconstruction.

Related Resources & Content

  1. npj Digital Medicine, 2025 -- Transforming Paper ECGs into Digital Format: An Open-Source Algorithm for Research in Clinical Settings
  2. npj Digital Medicine, 2025 -- The cost of explainability in artificial intelligence-enhanced electrocardiogram models
  3. npj Digital Medicine, 2025 -- Methods of ECG Sonification for Enhanced and Versatile Clinical Decision-Making Support
  4. 2025 ACC/AHA/ACEP/NAEMSP/SCAI Guideline for the Management of Patients With Acute Coronary Syndromes
  5. Reconstruction of 12-lead ECG: a review of algorithms - PMC
  6. A Comprehensive Deep Learning Strategy for Detecting Coronary Artery Calcifications in 2D Angiographic Images
  7. 2025 ACC/AHA/ACEP/NAEMSP/SCAI Guideline for the Management of Patients With Acute Coronary Syndromes: A Report of the American College of Cardiology/American Heart Association Joint Committee on Clinical Practice Guidelines
  8. IEC 60601-2-25 — Article List — MEDTEQ
  9. Reconstruction of 12-lead ECG: a review of algorithms - PMC

Original Source(s)

Related Content