RECOGNIZING FACIAL EXPRESSIONS

1
RECOGNIZING FACIAL EXPRESSIONS THROUGH TRACKING
Salih Burak Gokturk
2
OVERVIEW

• PROBLEM DESCRIPTION
• TRAINING STAGE
• TESTING STAGE
• EXPERIMENTS
• CONCLUSION

3
Components of the recognition system
Training with stereo
Data
Classifier
Testing with mono
New Data
Output
Analysis -Face Tracking

• Intelligence
• Support Vector Machine
• Classifier

Shape Parameters
4
PROBLEM DESCRIPTION(Tracking )
5
PROBLEM DESCRIPTION (Recognition)
Training
Testing
Data
Classifier
New Data
Output
6
OVERVIEW

• PROBLEM DESCRIPTION
• TRAINING STAGE
• TESTING STAGE
• EXPERIMENTS
• CONCLUSION

7
Monocular Tracking And Classification
Stereo Tracking
Learn Shape
Data

• p - degrees of freedom

8
Support Vector Machines (SVM)
Training
Testing (Classifier)
Data
Classifier
New Data
Output
- Best discriminating hypersurface between
two class of objects

• Map the data to high dimension
• using a map function ?

• The hypersurface in the feature
• space corresponds to a hyperplane
• in the mapped space

9
OVERVIEW

• PROBLEM DESCRIPTION
• TRAINING STAGE
• TESTING STAGE
• EXPERIMENTS
• CONCLUSION

10
LUKAS TOMASI KANADE OPTICAL FLOW TRACKER EXTENDED
TO 3D
X(t1)
X(t)
?
TIME t1
I(t1)
11
One to Many Application of Support Vector
Machines (SVM)
- One hypersurface per class is calculated
- A new data is tested for each hypersurface
- A different probability is assigned to ith class
12
OVERVIEW

• PROBLEM DESCRIPTION
• TRAINING STAGE
• TESTING STAGE
• EXPERIMENTS
• CONCLUSION

13

• Training (Stereo) with 2 people, totally 240
  frames
• Testing with 3 people
• 5 expressions neutral, open mouth, close mouth,
  
• smile, raise eyebrow
• velocity term is added to the shape vector

• Two other classifiers were tested
• 1 - Clustering 2 N-Nearest Neighbor

14
MOVIE (1)
15
MOVIE (2)
16
Performance of the system for different
expressions
Table 1
17
Table 2
18

• Training (Stereo) with 1 person, totally 130
  frames
• Testing with 3 people
• 5 expressions neutral, open mouth, close mouth,
  
• smile, raise eyebrow

Table 3
19

• Training (Stereo) with 2 people, totally 240
  frames
• Testing with 3 people
• 3 emotional expressions neutral, happy,
  surprise
• Transition between expressions are separated

Table 4
20
(No Transcript)
21
OVERVIEW

• PROBLEM DESCRIPTION
• TRAINING STAGE
• TESTING STAGE
• EXPERIMENTS
• CONCLUSION

22
Conclusions

• Breakthrough facial expression recognition rates
  .
• 3-D is the right way to go

Future Work

• Test with more subjects and expressions.
• further application to face recognition (?)