Related Experiment Video
Updated: Jul 5, 2025
04:48
Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
Published on: November 30, 2022
2.8K
Automatic Segmentation of Facial Regions of Interest and Stress Detection Using Machine Learning.
Daniel Jaramillo-Quintanar1, Jean K Gomez-Reyes1, Luis A Morales-Hernandez1
1Laboratory of Artificial Vision and Thermography/Mechatronics, Faculty of Engineering, Autonomous University of Queretaro, Campus San Juan del Rio, San Juan del Rio 76807, Mexico.
Sensors (Basel, Switzerland)
|January 11, 2024
Summary
This study introduces a novel machine learning method to detect human stress using facial thermal imaging. The system accurately classifies stress states, offering a non-invasive tool for stress assessment.
Area of Science:
- Thermography
- Machine Learning
- Human Physiology
Background:
- Stress significantly impacts quality of life, necessitating reliable detection methods.
- Current stress assessment tools can be invasive or complex.
- Non-invasive, user-friendly stress detection is crucial for widespread application.
Purpose of the Study:
- To develop and validate a machine learning methodology for automatic human stress classification.
- To utilize facial thermal imaging for non-invasive stress detection.
- To create a robust tool for identifying stress, baseline, and relaxed states.
Main Methods:
- Automatic detection of facial regions of interest (nose, cheeks, forehead, chin) in thermal images.
- Extraction of temperature data from these regions.
- Classification of stress states using a Support Vector Machine (SVM) classifier.
- Testing on 25 participants undergoing the Trier Social Stress Test and relaxation.
Main Results:
- Achieved a classification accuracy of 95.4%.
- Recorded a low error rate of 4.5%.
- Demonstrated successful differentiation between baseline, stressed, and relaxed states.
Conclusions:
- The proposed methodology enables automatic stress state classification from facial thermal images.
- This innovative tool is suitable for specialists and online applications.
- The system offers a robust and accurate approach to non-invasive stress monitoring.
