Non-contact blood pressure estimation by imaging photoplethysmography

Area of Science:

• Biomedical Engineering
• Medical Imaging
• Physiology

Background:

• Continuous, non-contact blood pressure (BP) monitoring is crucial for clinical and out-of-hospital applications.
• Current methods for BP estimation using imaging photoplethysmography (iPPG) are limited in scope and understanding.
• Further research is needed to optimize processing strategies and assess the general applicability of iPPG for BP estimation.

Purpose of the Study:

• To evaluate the quality of continuous BP estimation via iPPG.
• To quantify the impact of various video processing and signal formation strategies on BP estimation accuracy.
• To enhance understanding of the mechanisms underlying successful and unsuccessful iPPG-based BP estimation.

Main Methods:

• A feature-based approach was employed for continuous BP estimation.
• Experimental data from 43 healthy subjects under different stimuli were used for training and testing.
• Leave-one-subject-out cross-validation was performed to assess model generalizability.

Main Results:

• Minimum median mean absolute errors (MAE) achieved were 12.8 mmHg (SBP), 8.81 mmHg (DBP), and 7.14 mmHg (PP).
• Maximum median correlation coefficients were 0.25 (SBP), 0.70 (DBP), and 0.79 (PP).
• Advanced video processing (Level Sets) and color channel combinations improved DBP and PP estimation compared to a baseline.

Conclusions:

• iPPG shows potential for non-contact BP estimation, particularly for DBP and PP, with improved processing methods.
• While correlations for DBP and PP are promising, absolute errors necessitate further research and development for clinical viability.
• Basic research into iPPG-based BP estimation is essential to overcome current limitations and enable future applications.