Non-Contact Heart Rate Detection When Face Information Is Missing during Online Learning.

Area of Science:

• Physiological signal analysis
• Biomedical engineering
• Human-computer interaction

Background:

• Physiological signals offer objective emotional analysis.
• Non-contact heart rate monitoring (rPPG) is used in various fields.
• Traditional rPPG requires static, frontal faces, limiting real-world application.

Purpose of the Study:

• To develop a robust non-contact heart rate detection algorithm.
• To overcome limitations of traditional rPPG methods regarding head rotation and partial face visibility.
• To enable practical heart rate monitoring in scenarios like online learning.

Main Methods:

• Developed a symmetry substitution method to handle undetected regions of interest (ROIs).
• Utilized face-eye location to determine ROIs when face information is partially missing.
• Algorithm designed to address head deflections between 30 and 45 degrees.

Main Results:

• Successfully addressed practical challenges in heart rate detection.
• Achieved a root mean square error (RMSE) below 7.64 bpm.
• Demonstrated improved performance in scenarios with partial face visibility and head rotation.

Conclusions:

• The proposed method enhances the practicality of remote photoplethysmography.
• This approach expands the applicability of non-contact heart rate monitoring.
• The algorithm offers a viable solution for heart rate detection in challenging real-world conditions.