DeScoD-ECG: Deep Score-Based Diffusion Model for ECG Baseline Wander and Noise Removal

Case ID:

The DeScoD-ECG model is a novel electrocardiogram (ECG) baseline wander and noise removal technology that extends the current diffusion model for reconstructing electrocardiogram signals. It establishes a new benchmark in biomedical signal processing to provide a more accurate approximation of ECG signals for detection of cardiovascular irregularities. The DeSCOD-ECG model shows overall improvement on conducted experiments over traditional digital filter-based and deep learning-based methods, leading to better approximations of the true data distribution and higher stability under extreme noise corruption.

Electrocardiogram signals commonly suffer noise interference such as baseline wander. High-quality and high-fidelity reconstruction of the ECG signals is of great significance to diagnosing cardiovascular diseases. This technology extends the conditional diffusion-based generative model for ECG noise removal. 


  • ECG signal processing 
  • Biomedical engineering
  • Cardiovascular disease detection
  • Telecommunications (noise reduction and signal enhancement)


  • More stable and consistent with different noise levels
  • Better approximations of data
  • Experimental results demonstrate state-of-the-art performance
Patent Information:
Contact For More Information:
Jay Martin
Licensing Associate, Software and Copyright
The University of Arizona
Lead Inventor(s):
Ao Li
Huayu Li
Gregory Ditzler
Janet Roveda