Scribd
Upload
187 views

Case Study Pca1-1

The document describes the eigenfaces approach for face recognition and detection. It involves representing face images as vectors, computing their average, and projecting the face vectors into a lower-dimensional "eigenface" space based on the eigenvectors of the covariance matrix of the face vectors. Faces can be recognized by projecting a test face into this space and finding the closest face vector, while faces can be detected by thresholding the distance from this space. Experiments showed the approach is robust to lighting but performance decreases with orientation and size changes.

Uploaded by

abdulbarimalik
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
187 views

Case Study Pca1-1

The document describes the eigenfaces approach for face recognition and detection. It involves representing face images as vectors, computing their average, and projecting the face vectors into a lower-dimensional "eigenface" space based on the eigenvectors of the covariance matrix of the face vectors. Faces can be recognized by projecting a test face into this space and finding the closest face vector, while faces can be detected by thresholding the distance from this space. Experiments showed the approach is robust to lighting but performance decreases with orientation and size changes.

Uploaded by

abdulbarimalik
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 11

Eigenfaces for Face Detection/Recognition

(M. Turk and A. Pentland, "Eigenfaces for Recognition", Journal of Cognitive Neuroscience, vol. 3, no. 1, pp. 71-86, 1991, hard copy)

Face Recognition
- The simplest approach is to think of it as a template matching problem:

- Problems arise when performing recognition in a high-dimensional space. - Signicant improvements can be achieved by rst mapping the data into a lowerdimensionality space. - How to nd this lower-dimensional space?

Main idea behind eigenfaces


- Suppose is an N 2 x1 vector, corresponding to an N x N face image I . - The idea is to represent (= - mean face) into a low-dimensional space:

mean = w1 u1 + w2 u2 + . . . w K u K ( K << N 2 )

-2-

Computation of the eigenfaces


Step 1: obtain face images I1 , I2 , ..., I M (training faces) (very important: the face images must be centered and of the same size)

Step 2: represent every image I i as a vector i Step 3: compute the average face vector :

=
Step 4: subtract the mean face:

1 M

i i=1

i = i
Step 5: compute the covariance matrix C :

1 C= M

n T n n=1

= AAT ( N 2 x N 2 matrix)
( N 2 x M matrix)

where A = [1 2 . . . M ]

-3Step 6: compute the eigenvectors ui of AAT The matrix AAT is very large --> not practical !! Step 6.1: consider the matrix AT A ( M x M matrix) Step 6.2: compute the eigenvectors v i of AT A

AT Av i = i v i
What is the relationship between us i and v i ?

AT Av i = i v i => AAT Av i = i Av i => CAv i = i Av i or Cui = i ui where ui = Av i


Thus, AAT and AT A have the same eigenvalues and their eigenvectors are related as follows: ui = Av i !! Note 1: AAT can have up to N 2 eigenvalues and eigenvectors. Note 2: AT A can have up to M eigenvalues and eigenvectors. Note 3: The M eigenvalues of AT A (along with their corresponding eigenvectors) correspond to the M largest eigenvalues of AAT (along with their corresponding eigenvectors). Step 6.3: compute the M best eigenvectors of AAT : ui = Av i (important: normalize ui such that ||ui|| = 1) Step 7: keep only K eigenvectors (corresponding to the K largest eigenvalues)

-4-

Representing faces onto this basis


- Each face (minus the mean) i in the training set can be represented as a linear combination of the best K eigenvectors:

i mean =

w ju j, j =1

( w j = uT j i )

(we call the u j s eigenfaces)

- Each normalized training face i is represented in this basis by a vector:


i w1 i w2 i = , i = 1, 2, . . . , M ... i w K

-5-

Face Recognition Using Eigenfaces


- Given an unknown face image (centered and of the same size like the training faces) follow these steps: Step 1: normalize : = Step 2: project on the eigenspace

wi ui i=1

( w i = uT i )

w1 w2 Step 3: represent as: = ... wK


Step 4: nd er = minl || l|| Step 5: if er < T r , then is recognized as face l from the training set. - The distance er is called distance within the face space (difs) Comment: we can use the common Euclidean distance to compute er , however, it has been reported that the Mahalanobis distance performs better:

|| || =
k

(variations along all axes are treated as equally signicant)

i=1

1
i

( w i w ik )2

-6-

Face Detection Using Eigenfaces


- Given an unknown image Step 1: compute =

= Step 2: compute

wi ui i=1

( w i = uT i )

|| Step 3: compute e d = ||
Step 4: if e d < T d , then is a face. - The distance e d is called distance from face space (dffs)

-7- Reconstruction of faces and non-faces

-8-

Time requirements
- About 400 msec (Lisp, Sun4, 128x128 images)

Applications
- Face detection, tracking, and recognition

Problems
- Background (deemphasize the outside of the face, e.g., by multiplying the input image by a 2D Gaussian window centered on the face) - Lighting conditions (performance degrades with light changes) - Scale (performance decreases quickly with changes to the head size) * multiscale eigenspaces * scale input image to multiple sizes) - Orientation (perfomance decreases but not as fast as with scale changes) * plane rotations can be handled * out-of-plane rotations more difcult to handle

-9-

Experiments
- 16 subjects, 3 orientations, 3 sizes - 3 lighting conditions, 6 resolutions (512x512 ... 16x16) - Total number of images: 2,592

-10Experiment 1 * Used various sets of 16 images for training * One image/person, taken under the same conditions * Eigenfaces were computed ofine (7 eigenfaces were used) * Classify the rest images as one of the 16 individuals * No rejections (i.e., no threshold for difs)

- Performed a large number of experiments and averaged the results: 96% correct averaged over light variation 85% correct averaged over orientation variation 64% correct averaged over size variation

-11Experiment 2 - They considered rejections (i.e., by thresholding difs) - There is a tradeoff between correct recognition and rejections. - Adjusting the threshold to achieve 100% recognition acurracy resulted in: * 19% rejections while varying lighting * 39% rejections while varying orientation * 60% rejections while varying size

Experiment 3 - Reconstruction using partial information

You might also like