Neck-focused Remote Photoplethysmography (rPPG): A comparative study using clinical data and the PyVHR framework

Area of Science:

• Biomedical Engineering
• Medical Imaging
• Physiological Monitoring

Background:

• Remote Photoplethysmography (rPPG) enables non-invasive physiological measurements using digital cameras.
• Current rPPG methods primarily focus on the face, limiting data privacy for subjects with medical conditions.
• The neck presents a viable alternative region of interest for rPPG due to its limited identifiable characteristics.

Purpose of the Study:

• To evaluate the feasibility of neck-based rPPG for accurate heart rate measurement.
• To compare the performance of neck-based rPPG with traditional face-based rPPG algorithms.
• To investigate the impact of subject posture on the accuracy of neck-based rPPG.

Main Methods:

• Utilized a modified PyVHR framework to process both neck and face videos.
• Applied 20 different rPPG algorithms to assess heart rate measurement accuracy.
• Introduced and employed the Heart-Rate Within Tolerance (HRWT) metric to evaluate reliability.

Main Results:

• Achieved comparable heart rate measurement accuracy on neck videos versus face videos (1.79 ± 4.93% BPM difference in mean absolute percentage error).
• Neck-based rPPG demonstrated an improved HRWT metric, outperforming facial videos by an average of 15%.
• Subject posture significantly affected neck-based rPPG accuracy, with reduced performance in seated versus supine positions.

Conclusions:

• Neck-based rPPG is a promising, privacy-preserving alternative for remote heart rate monitoring.
• Neck-based rPPG achieves comparable accuracy to face-based methods and offers improved reliability.
• Further research should consider subject posture optimization for neck-based rPPG applications.