Identification and Tracking of Physiological Parameters from Skin using Video Photoplethysmography.

Area of Science:

• Biomedical Engineering
• Physiological Monitoring
• Signal Processing

Background:

• Video Photoplethysmography (vPPG) estimates cardiovascular parameters from skin video recordings.
• Challenges include motion artifacts, lighting changes, and accurate skin-background differentiation.
• Existing vPPG methods struggle with reliable blood volume pulse waveform extraction.

Purpose of the Study:

• To develop and validate a robust skin detection algorithm for vPPG.
• To improve the accuracy of cardiovascular parameter estimation from video.
• To assess the algorithm's performance under varying conditions, including thermal stimulation.

Main Methods:

• An algorithm was developed to detect skin pixels robust to luminosity variations.
• Video recordings were acquired during a thermal stimulation protocol on 10 healthy subjects.
• Processed vPPG signals were analyzed using band-pass filtering, peak detection for heart rate variability (HRV), and compared to finger-PPG.

Main Results:

• The vPPG algorithm demonstrated statistical agreement with finger-PPG in time and frequency domain HRV indexes.
• The proposed skin detection method showed stability against luminosity changes.
• Preliminary findings indicate good overall performance, with room for improvement in frequency component estimation.

Conclusions:

• The developed skin detection algorithm enhances vPPG signal quality and cardiovascular parameter estimation.
• The study validates vPPG as a viable, non-invasive technique for physiological monitoring.
• Further research is needed to optimize frequency-domain analysis, especially during resting conditions.