Related Experiment Videos
Facial action recognition for facial expression analysis from static face images
Maja Pantic1, Leon J M Rothkrantz
1Department of Electrical Engineering, Mathematics and Computer Science, Delft University of Technology, 2628 CD Delft, The Netherlands. M.Pantic@ewi.tudelft.nl
Summary
This study introduces an automated system for recognizing facial gestures from static face images. The system accurately identifies 32 facial muscle actions with an 86% recognition rate.
Area of Science:
- Computer Vision
- Biometrics
- Human-Computer Interaction
Background:
- Facial gesture recognition is a growing area in machine vision.
- Automated analysis of facial muscle activity is crucial for various applications.
Purpose of the Study:
- To develop an automated system for recognizing facial gestures in static, frontal, and profile view color face images.
- To identify individual facial muscle actions (AUs) and their combinations.
Main Methods:
- Utilized a multidetector approach for facial feature localization.
- Extracted fiducial points from facial contours (profile, eyes, mouth).
- Employed rule-based reasoning to recognize 32 facial muscle actions (AUs) with certainty factors.
Main Results:
- Achieved an 86% recognition rate for facial muscle actions.
- Successfully recognized individual AUs and their combinations.
Conclusions:
- The developed automated system demonstrates high accuracy in facial gesture recognition.
- This system has potential applications in areas requiring facial expression analysis.