Key result
Camera-based PTT correlates better with systolic BP while red-channel AIx tracks diastolic BP.
Why the study?
Hemodynamic features and waveform features used for camera-based photoplethysmography noncontact blood pressure estimation had not been rigorously investigated or benchmarked.
Do specific hemodynamic and waveform features from camera-based photoplethysmography accurately correlate with blood pressure?
Observational (n=16)
Do specific hemodynamic and waveform features from camera-based photoplethysmography accurately correlate with blood pressure?
Camera-based blood pressure estimation can be optimized by using neck-to-forehead pulse transit time for systolic BP and red-channel augmented index for diastolic BP.
Optimizes feature selection for contactless blood pressure monitoring; leaves open whether these algorithms can achieve clinical-grade.
Blood pressure (BP) is a major vital sign that can indicate the cardiac function and peripheral vascular resistance of a subject. Recent studies demonstrated the feasibility of using camera-based photoplethysmography (camera-PPG) signals for noncontact BP measurement. Most prior arts exploited two types of PPG features for BP estimation: hemodynamic features [pulse transit time (PTT)] and waveform features (dicrotic notch). However, these two types of features have not been rigorously investigated nor benchmarked. In this article, we thoroughly investigated these features in the ice water stimulation experiment involving 16 subjects, by comparing: 1) the PTT from multisite PPG signals and the PTT from the camera pressure pulse wave (camera-PPW) signal from the neck and the PPG signal from the forehead; 2) the augmented index (AIx) measured from the PPG waveform from the forehead in different wavelength channels of an RGB camera; and 3) the hemodynamic feature (PTT) and waveform feature (AIx). The major findings are: the PTT measured between the neck-PPW signal and forehead-PPG signal has a better correlation with BP than the PTT measured from multisite PPG signals; the waveform feature AIx has wavelength-dependence, where the red channel with deeper skin penetration depth is more reliable for waveform-based BP estimation; the waveform feature AIx has a better correlation with diastolic BP while the PTT is better correlated with Systolic BP. The findings are understandable from the physiological perspective and the insights obtained from the study are useful for creating a noncontact BP measurement solution.
No takes yet. Share an insight, caveat, or question.
Liao et al. (2023) conducted an observational in Blood pressure estimation (n=16). Camera-based photoplethysmography (camera-PPG) features was evaluated on Correlation with systolic and diastolic blood pressure. Camera-based pulse transit time (PTT) from neck and forehead signals correlated better with systolic BP, while the augmented index (AIx) from the red channel correlated better with diastolic BP.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: