Generalized channel separation algorithms for accurate camera-based multi-wavelength PTT and BP estimation

Area of Science:

• Biomedical Engineering
• Optical Sensing
• Physiological Monitoring

Background:

• Multi-wavelength (MW) pulse transit time (PTT) enables continuous blood pressure (BP) monitoring but is challenged by skin compression in contact-based methods.
• Extending MW PTT to contact-free camera sensing is hindered by color channel overlap in RGB cameras, mixing wavelength responses and preventing accurate PTT measurement.

Purpose of the Study:

• To develop and validate novel camera-independent, data-driven channel separation algorithms for contact-free MW PTT and BP estimation.
• To overcome the limitations of color channel overlap in consumer RGB cameras for physiological signal acquisition.

Main Methods:

• Proposed novel camera-independent data-driven channel separation algorithms utilizing constrained genetic algorithms.
• Systematically validated algorithms on camera recordings of palms and ground-truth BP measurements from 13 subjects in rest and activity scenarios.
• Compared proposed algorithms against established blind source separation methods and a camera-specific physics-based method.

Main Results:

• The best-performing algorithm achieved Mean Absolute Errors (MAEs) of 3.48 mmHg for systolic BP and 2.61 mmHg for diastolic BP in a leave-one-subject-out experiment with personalization.
• Demonstrated good performance in both PTT reconstruction and BP estimation using a Random Forest regressor.
• The proposed algorithms effectively addressed color channel overlap, enabling meaningful PTT measurement from camera data.

Conclusions:

• The developed algorithms enable novel, contact-free MW PTT and BP estimation.
• This approach offers a promising avenue for non-invasive, continuous blood pressure monitoring using readily available camera technology.
• The data-driven channel separation method provides a robust solution for overcoming challenges in camera-based physiological sensing.