Wearable PPG-to-Multi-Lead ECG Conversion for Cardiac Monitoring

Chongxin Zhong, Zhishan Guo, Anil Kishin Gehi, Chenhan Xu, Huining Li

Confirmation: I have read and agree with the IEEE BSN 2025 conference submission's policy on behalf of myself and my co-authors.

Keywords: PPG, ECG, Cardiac Monitoring, Wearable

TL;DR: We present a wearable approach for converting PPG signals into multi-lead ECG data, which can support continuous and wearable cardiac monitoring.

Abstract:

The electrocardiogram (ECG) has been the gold standard for heart disease evaluation due to the rich information about the electrical activity of the heart contained in it. However, existing ECG monitoring devices either lack the capability for continuous monitoring or are unable to support multi-lead ECG recordings. To address the issues, we propose an approach for generating multi-lead ECG from photoplethysmogram (PPG), which can be passively monitored by wearable devices such as smartwatches. The PPG collected from wearable devices is first passed to a trained conditional diffusion model to generate the single-lead ECG, and then through a long
short-term memory (LSTM) model to construct and predict the multi-lead ECG. The final outputs can be used to monitor and detect abnormal cardiac patterns in daily life. We evaluate the performance of our proposed approach with the dataset collected from daily-life scenarios involving 32 subjects. The results show that our approach can generate multi-lead ECGs accurately. In addition, a case study is conducted using data collected from the hospital, which demonstrates the effectiveness of our approach in detecting ST elevation.

Track: 1. Digital Health Solutions (i.e. sensors and algorithms) for diagnosis, progress, and self-management

NominateReviewer: Chongxin Zhong, [e-mail]

Submission Number: 108