PaperarXiv:2605.12541
PG-LRF: Physiology-Guided Latent Rectified Flow for Electro-Hemodynamic PPG-to-ECG Generation
NeurIPS 2026 (listed on Emory CS page)
TL;DR
PG-LRF introduces physiology-guided latent rectified flow for PPG-to-ECG generation on MC-MED. Simulator-regularized flow improves waveform fidelity and interval metrics over prior generative baselines and lifts a downstream CVD label probe on synthesized ECG.
- arXiv 2605.12541
- NeurIPS 2026 listing on Emory lab page
- MC-MED waveform MAE 0.71 / HR MAE 3.94 bpm
- Interval MAE (ms): PR 59.9, QRS 38.6, QT 29.3, QTcF 32.2
- Generated-ECG CVD Macro-AUROC 0.633 vs PaPaGei-on-PPG 0.545
Results on ChatPPG leaderboards
All leaderboardsNo results from this paper on ChatPPG leaderboards yet. Browse the leaderboards.
Tasks & signals
Signals
Keywords
Models
How it works
Cite this paper
APA
Wang, X., Wang, M., Han, K., Cao, D., Chang, C., Shi, Y., Yan, R., Luo, X., Liu, Y., Hu, X., Sun, Y., Wang, W., & Yang, C. (2026). PG-LRF: Physiology-Guided Latent Rectified Flow for Electro-Hemodynamic PPG-to-ECG Generation. NeurIPS 2026. https://arxiv.org/abs/2605.12541
BibTeX
@inproceedings{wang2026pglrf,
title = {{PG-LRF: Physiology-Guided Latent Rectified Flow for Electro-Hemodynamic PPG-to-ECG Generation}},
author = {Wang, Xiaoda and Wang, Minxiao and Han, Kaiqiao and Cao, Defu and Chang, Ching and Shi, Yidan and Yan, Runze and Luo, Xiao and Liu, Yan and Hu, Xiao and Sun, Yizhou and Wang, Wei and Yang, Carl},
year = {2026},
booktitle = {NeurIPS 2026},
eprint = {2605.12541},
archivePrefix = {arXiv}
}Built from the hub record. Check it against the publisher before submitting.
Added to the ChatPPG research hub
