1. PART 1 Face Recognition and Its applications
Based on works of: Jinshan Tang; Ariel P from Hebrew University; Mircea Focşa, UMFT; Xiaozhen Niu,
Department of Computing Science, University of Alberta; Christine Podilchuk,

2. Contents Introduction Face detection using color information
Face matching
Face Segmentation/Detection
Facial Feature extraction
Face Recognition
Video-based Face Recognition
Comparison
Conclusion
Reference

3. Face Segmentation/Detection
During the past ten years, considerable progress has been made in multi-face recognition area,
This includes:
Example-based learning approach by Sung and Poggio (1994).
The neural network approach by Rowley et al. (1998).
Support vector machine (SVM) by Osuna et al. (1997).

4. Introduction

6. Basic steps for face recognition
Input face image
Face detection
Face recognition
Face feature
extraction
Face database
Feature Matching
Decision maker
Output result

7. Face detection Geometric information based face detection
Color information based face detection
Combining them together
(a) Geometric information based face detection
(b) Color information based face detection

8. Color information based face detection
Face color is different from background
Choice of color spaces is very important
Color Spaces:
R,G,B
YCbCr
YUV
r,g
……..
Skin
color
Background
color
Figure 4. Skin color distribution in a complex background

9. A face detection algorithm Using Color and Geometric information

10. Ideas: (1) compensate for lightning, (2) separate by transforming to new (sub) space.

11. Ideas: (1) compensate for lightning, (2) separate by transforming to new (sub) space.
(3) clustering.

12. Feature-based face detection
Color can be used in segmentation and grouping of image subareas.

13. Location and shape parameters of eyes are the most important features to be detected through segmentation and morphological operations (dilation and erosion).

15. Ideas:
Eyes
Mouth
Boundary (edge detection)
Boundary approximated to ellipse or something (Hough)

16. The concept of eye glasses
The concept of half-profiles

17. Face Matching Feature based face matching Template matching
Features versus templates

18. Feature based face matching
You can extract various features
Face image
From face
detection
Normalization
Feature extraction
Feature vector
classifier
Output results
Decision
maker
You can use various decision makers
You can use various classifiers

19. Normalization Eye location
Normalization: rotation normalization, scale normalization
Averaged for objects
Cross Correlation :
object
template

20. Feature extraction 3.5-D feature vector
Eyebrow thickness and vertical position at the eye center position
A coarse description of the left eyebrow’s arches
Nose vertical position and width
Mouth vertical position, width, height upper and lower lips
eleven radii describing the chin shape
Bigonial breadth (face width at nose position)
Zygomatic breadth (face width halfway between nose tip and eyes).
3.5-D feature vector

21. Example of some geometrical features

22. Classifier Bayes classifier Rank the distance values
This is just one example of classifier, others are Decision Trees, expressions, decomposed structures, NNs.
Bayes classifier
Rank the distance values
Output the results
Feature vector
Computer
(j=2,3,…N)

23. ANN Classifier ANN one-class-in-one network multi-class-in-one network
Feature vector
Class 1
Class 2
ANN
one-class-in-one network
multi-class-in-one network
MAXNET
Classification results
Fig.2. one-class-in-one network

24. Template matching Templates database Produce a template Face image
You have to create the data base of templates for all people you want to recognize
Produce a template
Face image
From face
detection
Normalization
matching
Output results
Decision
maker

25. There are different templates used in various regions of the normalized face.
Various methods can be used to compress information for each template.

26. Example-based learning approach (EBL)
Three parts:
The image is divided into many possible-overlapping windows,
each window pattern gets classified as either “a face” or “not a face” based on a set of local image measurements.
For each new pattern to be classified, the system computes a set of different measurements between the new pattern and the canonical face model.
A trained classifier identifies the new pattern as “a face” or “not a face”.

27. Example of a system using EBL

28. Neural network (NN)
Kanade et al. first proposed an NN-based approach in 1996.
Although NN have received significant attention in many research areas, few applications were successful in face recognition.
Why?

29. Neural network (NN)
It’s easy to train a neural network with samples which contain faces, but it is much harder to train a neural network with samples which do not.
The number of “non-face” samples are just too large.

30. Neural network (NN) Neural network-based filter.
A small filter window is used to scan through all portions of the image,
and to detect whether a face exists in each window.
Merging overlapping detections and arbitration. By setting a small threshold, many false detections can be eliminated.

31. An example of using NN

32. Test results of using NN

33. SVM (Support Vector Machine)
SVM was first proposed in 1997, it can be viewed as a way to train polynomial neural network or radial basic function classifiers.
Can improve the accuracy and reduce the computation.

34. Comparison with Example Based Learning (EBL)
Test results reported in 1997.
Using two test sets (155 faces).
SVM achieved better detection rate and fewer false alarms.

35. Recent approaches
Face segmentation/detection research area still remain active, for example:
An integrated SVM approach to multi-face detection and recognition was proposed in 2000.
A technique of background learning was proposed in August 2002.
Still lots of potential!

36. Static face recognition
Numerous face recognition methods/algorithms have been proposed in last 20 years,
several representative approaches are:
Eigenface
LDA/FDA (Linear DA, Fisher DA) Discriminant analysis (algorithm)
Neural network (NN)
PCA – Principal Component Analysis
Discrete Hidden Markov Models (DHMM)
Continuous Density HMM (CDHMM).

37. Eigenface The basic steps are:
Registration. A face in an input image first must be located and registered in a standard-size frame.
Eigenpresentation.
Every face in the database can be represented as a vector of weights,
the principal component analysis (PCA) is used to encode face images and capture face features.
Identification. This part is done by locating the images in the database whose weights are the closest (in Euclidean distance) to the weights of the test images.

38. LDA/FDA
Face recognition method using LDA/FDA is called the fishface method.
Eigenface use linear PCA. It is not optimal to discrimination for one face class from others.
Fishface method seeks to find a linear transformation to maximize the between-class scatter and minimize the within-class scatter.
Test results demonstrated LDA/FDA is better than eigenface using linear PCA (1997).

39. Test results of LDA
Test results of a subspace LDA-based face recognition method in 1999.

40. Video-based Face Recognition
Three challenges:
Low quality
Small images
Characteristics of face/human objects.
Three advantages:
Allows much more information.
Tracking of face image.
Provides continuity,
this allows reuse of classification information from high-quality images in processing low-quality images from a video sequence.

41. Basic steps for video-based face recognition
Object segmentation/detection.
Motion structure.
The goal of this step is to estimate the 3D depths of points from the image sequence.
3D models for faces.
Using a 3D model to match frontal views of the face.
Non-rigid motion analysis.

42. Recent approaches
Most video-based face recognition system has three modules for
detection,
tracking
and recognition.
An access control system using Radial Basis Function (RBS) network was proposed in 1997.
A generic approach based on posterior estimation using sequential Monte Carlo methods was proposed in 2000.
A scheme based on streaming face recognition (SFR) was propose in August 2002.

43. The Streaming Face Recognition (SFR) scheme
Combine several decision rules together, such as:
Discrete Hidden Markov Models (DHMM) and
Continuous Density HMM (CDHMM).
The test result achieved a 99% correct recognition rate in the intelligent room.

44. Comparison
Two most representative and important protocols for face recognition evaluations:
The FERET protocol (1994).
Consists of 14,126 images of 1199 individuals.
Three evaluation tests had been administered in 1994, 1996, and 1997.
The XM2VTS protocol (1999).
Expansion of previous M2VTS program (5 shots of each of 37 subjects).
Now consists 295 subjects.
The results of M2VTS/XM2VTS can be used in wide range of applications.

45. 1996/1997 FERET Evaluations
Compared ten algorithms.

46. Conclusion Face recognition has many potential applications.
For many years not very successful,
we need to improve the accuracy of face recognition
Combining face recognition and other biometric recognition technologies,
Such as:
fingerprint recognition technology,
voice recognition technologies
and so on
For our applications accuracy is much more important than speed.

47. Conclusion Significant achievements have been made recently.
LDA-based methods and NN-based methods are very successful.
FERET and XM2VTS have had a significant impact to the developing of face recognition algorithms.
Challenges still exist, such as pose changing and illumination changing.
Face recognition area will remain active for a long time.

48. Reference
[1] W. Zhao, R. Chellappa, A. Rosenfeld, and P.J. Phillips, Face Recognition: A Literature Survey, UMD CFAR Technical Report CAR-TR-948, 2000.
[2] K. Sung and T. Poggio, Example-based Learning for View-based Human Face Detection, A.I. Memo 1521, MIT A.I. Laboratory, 1994.
[3] H.A. Rowley, S. Baluja, and T. Kanade, Neural Network Based Face Detection, IEEE Trans. On Pattern Analysis and Machine Intelligence, Vol. 20, 1998.
[4] E. Osuna, R. Freund, and F. Girosi, Training Support Vector Machines: An Application to Face Recognition, in IEEE Conference on Computer Vision and Pattern Recognition, pp , 1997.
[5] M. Turk and A. Pentland, Eigenfaces for Recognition, Journal of Cognitive Neuroscience, Vol.3, pp , 1991.
[6] W. Zhao, Robust Image Based 3D Face Recognition, PhD thesis, University of Maryland, 1999.
[7] K.S. Huang and M.M. Trivedi, Streaming Face Recognition using Multicamera Video Arrays, 16th International Conference on Pattern Recognition (ICPR). August 11-15, 2002.
[8] P.J. Phillips, P. Rauss, and S. Der, FERET (Face Recognition Technology) Recognition Algorithm Development and Test Report, Technical Report ARL-TR 995, U.S. Army Research Laboratory.
[9] K. Messer, J. Matas, J. Kittler, J. Luettin, and G. Maitre, XM2VTSDB: The Extended M2VTS Database, in Proceedings, International Conference on Audio and Video-based Person Authentication, pp , 1999.