ECG Inpainting with Denoising Diffusion Prior

Abstract

In this work, we train a generative denoising diffusion model (DDGM) in healthy electrocardiogram (ECG) data capable of generating realistic healthy heartbeats. We then show how recent advances in solving linear inverse Bayesian problems with DDGM can be used to derive interpretable outlier detection tools for electrophysiological anomalies.

Cite

Text

Bedin et al. "ECG Inpainting with Denoising Diffusion Prior." NeurIPS 2023 Workshops: DGM4H, 2023.

Markdown

[Bedin et al. "ECG Inpainting with Denoising Diffusion Prior." NeurIPS 2023 Workshops: DGM4H, 2023.](https://mlanthology.org/neuripsw/2023/bedin2023neuripsw-ecg/)

BibTeX

@inproceedings{bedin2023neuripsw-ecg,
  title     = {{ECG Inpainting with Denoising Diffusion Prior}},
  author    = {Bedin, Lisa and Cardoso, Gabriel and Dubois, Remi and Moulines, Eric},
  booktitle = {NeurIPS 2023 Workshops: DGM4H},
  year      = {2023},
  url       = {https://mlanthology.org/neuripsw/2023/bedin2023neuripsw-ecg/}
}