Related Experiment Video
Updated: Aug 28, 2025
06:37
Author Spotlight: Addressing Technical and Subjective Challenges in Measuring Classroom Attention
Published on: December 15, 2023
4.0K
Using Computer Vision to Track Facial Color Changes and Predict Heart Rate
Salik Ram Khanal1,2, Jaime Sampaio1, Juliana Exel3
1Research Center in Sports Sciences, Health Sciences and Human Development, CIDESD, Universidade de Trás-os-Montes e Alto Douro, 5000-801 Vila Real, Portugal.
Journal of Imaging
|September 22, 2022
Summary
This study introduces a non-contact computer vision method to estimate exercise intensity by analyzing facial color changes. This approach offers a novel way to monitor physical exertion without requiring additional equipment.
Area of Science:
- Exercise Physiology
- Biomedical Engineering
- Computer Vision
Background:
- Monitoring exercise intensity is crucial for performance optimization and health management.
- Current methods often require specialized equipment, limiting accessibility.
- The relationship between facial color changes and exercise intensity remains underexplored.
Purpose of the Study:
- To develop a non-contact computer vision method for determining heart rate and exercise intensity.
- To investigate the efficacy of various color models (RGB, HSV, YCbCr, Lab, YUV) in assessing exercise intensity.
- To establish a reliable method for monitoring physical exertion without instrumentation.
Main Methods:
- Utilized computer vision techniques to analyze facial color variations during exercise.
- Employed multiple color models (RGB, HSV, YCbCr, Lab, YUV) for data analysis.
- Developed personalized and universal models using multiple auto regressions and polynomial regression to predict maximum heart rate percentage (maxHR%).
Main Results:
- The multiple polynomial regression model, using data from all participants, demonstrated the highest accuracy.
- This universal model achieved a Root Mean Square Error (RMSE) of 6.75 and an R-square value of 0.78.
- The study successfully established a correlation between facial color changes and exercise intensity.
Conclusions:
- Facial color analysis via computer vision presents a viable non-contact method for monitoring exercise intensity.
- This technology can enhance online monitoring systems for exercise prescription and control.
- The findings suggest a potential for widespread application in fitness and sports science without the need for external devices.
