Face Recognition

under the guidance of

Mrs. S. INDU

VIKAS PHOGAT 2K4/EC/684

YOGESH KUMAR 2K4/EC/691
RAHUL AHLAWAT 2K4/EC/662
 Face Recognition
 Introduction
 Face recognition algorithms
 Comparison
 Short summary of the presentation

2
 Introduction
 Why we are interested in face recognition?
 Passport control at terminals in airports
 Participant identification in meetings
 System access control
 Scanning for criminal persons

3
 Face Recognition Algorithms
 Types of FACE RECOGNITION ALGOS
 Eigenfaces
 Fisherfaces

4
 Eigenfaces
 Developed in 1991 by M.Turk
 Based on PCA
 Relatively simple
 Fast

5
 Eigenfaces
 PCA seeks directions that are efficient for
representing the data

not efficient efficient

Class A Class A

Class B Class B

6
 Eigenfaces
 PCA maximizes the total scatter

scatter

Class A

Class B

7
 Eigenfaces
 PCA reduces the dimension of the data
 Speeds up the computational time

8
 Eigenfaces, the algorithm
 Assumptions
 Square images with W=H=N
 M is the number of images in the database
 P is the number of persons in the database

9
 Eigenfaces, the algorithm
 The database
 a1   b1   c1   d1 
       
a2  b c d
      
2 2 2
      
 
       
a b c d
 N 
 N 
2
 N 2
 N 2 2

 e1   f1   g1   h1 
       
e f2  g h

2 
 
2 
 
2

        
       
e  f N 2  g h
 N2   N 2
 N 2

10
 Eigenfaces, the algorithm
 We compute the average face

 a1  b1    h1 
 
 1  a2  b2    h2 
m , where M  8
M   
 
 aN 2  bN 2    hN 2 

11
 Eigenfaces, the algorithm
 Then subtract it from the training faces
 a1  m1   b1  m1   c1  m1   d1  m1 
    b2  m2    c2  m2    d 2  m2 
  a2  m2 
am  , bm  , cm  , dm  ,
              
       
a
 N2  m N2 
b
 N2  m N2 
c
 N2  m N2 
d
 N2  m N2 

 e1  m1   f1  m1   g1  m1   h1  m1 
        
e  m f 2  m2  g  m  h  m
em  
2 
, gm  
2 
, hm  
 2  2 2 2 
, fm 
            
       
 eN 2  mN 2   f N 2  mN 2   g N 2  mN 2   hN 2  m N 2 

12
 Eigenfaces, the algorithm
 Now we build the matrix which is N2 by M
       
A   am bm cm d m em f m g m hm 

 The covariance matrix which is N2 by N2


Cov  AA

13
 Eigenfaces, the algorithm
 Find eigenvalues of the covariance matrix
 The matrix is very large
 The computational effort is very big

 We are interested in at most M eigenvalues

 We can reduce the dimension of the matrix

14
 Eigenfaces, the algorithm
 Compute another matrix which is M by M

LA A
 Find the M eigenvalues and eigenvectors
 Eigenvectors of Cov and L are equivalent

 Build matrix V from the eigenvectors of L

15
 Eigenfaces, the algorithm
 Eigenvectors of Cov are linear combination
of image space with the eigenvectors of L

U  AV
 Eigenvectors represent the variation in the
faces

16
 Eigenfaces, the algorithm
 Compute for each face its projection onto
the face space
   
   

1  U  am  ,  2  U 

bm , 3  U  cm  ,  4  U  dm ,
 
 
5  U  em  , 6  U     
f m ,  7  U  g m  , 8  U  hm

 Compute the threshold

1

  max i   j
2
 for i, j  1.. M

17
 Eigenfaces, the algorithm
 To recognize a face
 r1 
 
r2
 
 
 
 rN 2 

 Subtract the average face from it

 r1  m1 
 
  2 r  m2 
rm 
  
 
 rN 2  mN 2  18
 Eigenfaces, the algorithm
 Compute its projection onto the face space
 
  U  rm 

 Compute the distance in the face space

between the face and all known faces
2
    i
i
2
for i  1.. M

19
 Eigenfaces, the algorithm
 Reconstruct the face from eigenfaces

s  U
 Compute the distance between the face
and its reconstruction
 2
  rm  s
2

20
 Eigenfaces, the algorithm
 Distinguish between
 If    then it’s not a known face

 If ξ<θ then it’s a known face

21
 Eigenfaces, the algorithm
 Problems with eigenfaces
 Different illumination
 Different head pose
 Different alignment
 Different facial expression

22
 Fisherfaces
 Developed in 1997 by P.Belhumeur et al.
 Based on Fisher’s LDA
 Faster than eigenfaces, in some cases
 Has lower error rates
 Works well even if different illumination
 Works well even if different facial express.

23
 Fisherfaces
 LDA maximizes the between-class scatter
 LDA minimizes the within-class scatter

Class A
Class B

24
 Comparison

best ID rate: eigenfaces 80.0%, fisherfaces 93.2%

25
 Summary
 The algorithm introduced
 Eigenfaces
 Reduce the dimension of the data from N2 to M
 Fast recognition of faces
 Problems with illumination, head pose etc

26
 References
[1] M. Turk, A. Pentland, “Face Recognition Using Eigenfaces”
[2] J. Ashbourn, Avanti, V. Bruce, A. Young, ”Face Recognition Based on Symmetrization
and Eigenfaces”
[3] http://www.markus-hofmann.de/eigen.html
[4] P. Belhumeur, J. Hespanha, D. Kriegman, “Eigenfaces vs Fisherfaces: Recognition
using Class Specific Linear Projection”
[5] R. Duda, P. Hart, D. Stork, “Pattern Classification”, ISBN 0-471-05669-3, pp. 121-124
[6] F. Perronin, J.-L. Dugelay, “Deformable Face Mapping For Person Identification”, ICIP
2003, Barcelona

27
 End

Thank you for your attention

28