Effect of Observation Time and Estimation Interval on rPPG-based Blood Pressure Estimation

Key Points

• Blood pressure estimation becomes more precise as observation time increases, and estimation intervals are optimized—leading to reduced absolute errors.
• The average absolute error decreases when observation periods are sufficiently long and estimation intervals are not overly short, enhancing estimation performance.
• Using a dataset constructed at Leipzig University Hospital, rPPG signals are processed to optimize blood pressure estimates while utilizing recurrent neural networks.
• The findings highlight a significant reduction in performance differences between rPPG and contact-based PPG estimation methods through careful design of observation and estimation intervals.

Structured PICO

Abstract