Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

Measuring facial expressions by computer image analysis.

M S Bartlett1, J C Hager, P Ekman

  • 1Department of Cognitive Science, University of California, San Diego, USA. marni@salk.edu

Psychophysiology
|April 9, 1999
PubMed
Summary
This summary is machine-generated.

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Immunogenicity of subcutaneous TNF inhibitors and its clinical significance in real-life setting in patients with spondyloarthritis.

Rheumatology international·2021
Same author

Minor effect of loss to follow-up on outcome interpretation in the Swedish spine register.

European spine journal : official publication of the European Spine Society, the European Spinal Deformity Society, and the European Section of the Cervical Spine Research Society·2019
Same author

Long-term outcome after spinal fusion for isthmic spondylolisthesis in adults.

The spine journal : official journal of the North American Spine Society·2018
Same author

Three out of four disease-modifying anti-rheumatic drug-naïve rheumatoid arthritis patients meet 28-joint Disease Activity Score remission at 12 months: results from the FIN-ERA cohort.

Scandinavian journal of rheumatology·2017
Same author

Multilayer Architectures for Facial Action Unit Recognition.

IEEE transactions on systems, man, and cybernetics. Part B, Cybernetics : a publication of the IEEE Systems, Man, and Cybernetics Society·2012
Same author

Standardization in pathologic assessment of transitional cell cancer of the bladder.

Urologic oncology·2011
Same journal

What's in a Mean? Comparing Interbeat Interval Averaging Methods Across Variability Levels and Window Lengths.

Psychophysiology·2026
Same journal

Model-Free and Model-Based Learning in Human Fear Conditioning.

Psychophysiology·2026
Same journal

Examining the Impact of Acute Exercise and Arousal Reappraisal on Stressor-Evoked Psychological and Cardiovascular Responses.

Psychophysiology·2026
Same journal

Respiratory Sinus Arrhythmia and Hierarchical Dimensions of Psychopathology.

Psychophysiology·2026
Same journal

Probing Prediction-Related Processes in Language Using an EEG Word Stem Completion Paradigm.

Psychophysiology·2026
Same journal

sBOSC: A Method for Source-Level Identification of Neural Oscillations in Electromagnetic Brain Signals.

Psychophysiology·2026
See all related articles

Automated facial action detection using computer vision accurately identifies emotions. This novel system matches expert performance, enhancing behavioral research accessibility.

Area of Science:

  • Behavioral Science
  • Computer Vision
  • Affective Neuroscience

Background:

  • Facial expressions are key indicators of emotion, cognition, and social interaction.
  • The Facial Action Coding System (FACS) quantifies facial movements objectively.
  • Automating FACS analysis is crucial for wider research application.

Purpose of the Study:

  • To develop and evaluate an automated system for detecting facial actions using computer image analysis.
  • To compare the performance of different computer vision approaches for facial action unit detection.
  • To assess the accuracy of the automated system against human expert and non-expert performance.

Main Methods:

  • Applied computer image analysis to sequences of facial images.
  • Compared holistic spatial analysis, explicit feature measurement (e.g., wrinkles), and motion flow fields.

Related Experiment Videos

  • Developed a hybrid system combining these methods for classification.

    • Main Results:

    • The hybrid system achieved 91% accuracy in classifying six upper facial actions.
    • Outperformed human non-experts in facial action detection accuracy.
    • Matched the performance of highly trained FACS experts.

    Conclusions:

    • Automated facial action detection is feasible and highly accurate.
    • This technology can significantly broaden the accessibility of facial expression measurement as a research tool.
    • Potential applications include behavioral science and neuroscience research on emotion.