Heart rate estimation from color video sequences with high SNR

Area of Science:

• Biomedical Engineering
• Medical Imaging
• Physiological Monitoring

Background:

• Non-invasive heart rate (HR) monitoring is crucial for health surveillance.
• Existing methods may require specialized equipment or direct contact.
• Remote photoplethysmography (rPPG) offers a promising contactless approach.

Purpose of the Study:

• To develop and validate a novel method for estimating human heart rate (HR) from color videos of lip image sequences.
• To enhance signal quality and accuracy in remote HR estimation.
• To explore the potential for public health monitoring using conventional cameras.

Main Methods:

• Proposed the absorption intensity heartbeat modulation-averaged shifted histogram (AIHM-ASH) method.
• Utilized AIHM generated by heartbeat events for HR signal demodulation.
• Calculated instantaneous HR modulation depth from relative red blood cell (RBC) concentration in green channel images.
• Employed the ASH algorithm to suppress background noise and improve signal-to-noise ratio (SNR).

Main Results:

• The AIHM-ASH method demonstrated effective HR estimation in various experimental setups.
• Validated accuracy reached 99.23% with a root mean square error (RMSE) of 0.8 bpm.
• The ASH algorithm significantly improved SNR and reduced interference from tissue and veins.

Conclusions:

• The AIHM-ASH method provides an accurate and effective approach for non-invasive HR estimation using lip videos.
• This technique holds significant potential for advancing remote health monitoring and disease prevention.
• The use of conventional color video cameras makes this method widely applicable in public settings.