Title: 4EyesFace-Realtime face detection, tracking, alignment and recognition
14EyesFace-Realtime face detection, tracking,
alignment and recognition
- Changbo Hu, Rogerio Feris and Matthew Turk
2Overview
- Introduction
- Face Detection and Pose tracking
- Face Alignment
- Face Recognition
- Conclusions
3Introduction
Detection
Pose tracking
Alignment
Recognition
4Introduction
- Why this is a difficult problem?
- Facial Expressions, Illumination Changes, Pose,
etc. - Object
- Develop a fully automatic system, suitable for
real-time applications to locate and track human
faces, then to align and recognize the face. - Evaluate it on a large dataset.
5Face Detection
- Viola and Jones, 2001
- Simple features, which can be computed very
fast. - A variant of Adaboost is used both to select
the features and to train the classifier. - Classifiers are combined in a cascade which
allows background regions of the image to be
quickly discarded.
6Face detection
7Pose tracking
Based on Kentaro Toyamas IFA framework
8Face Alignment
- Active Appearance Model (AAM)
Statistical Shape Model (PCA)
Statistical Texture Model (PCA)
9Face alignment
- Problem Partial Occlusion
- Active Wavelet Networks (AWN) (on BMVC03)
- Main idea Replace AAM texture model by a
wavelet network
10Face Alignment
Similar performance to AAM in images under normal
conditions.
More robust against partial occlusions.
11Face Alignment
Using 9 wavelets, the system requires only
3 ms per iteration. In general, at most 10
iterations are sufficiently for good convergence
(PIV 1.6Ghz).
12Multi-View Face Alignment
- View selection by pose tracker
13Multi-View Face Alignment
14Face recognition
- online recognition
- HMM based face recognition
-
15Face recognition
- Large dataset evaluation
- FERET DataSet
- 1196 different individuals
- With ground truth of eye corners
16Face recognition
17Face recognition
18Face Recognition
19Face Recognition
20Conclusion
- We develop a system to do human face detection,
tracking, alignment and recognition - In this system, we invented new methods AWN and
extent to multi-view AWN - We implement the related detection and pose
tracking - Evaluate our method on large dataset