+ All Categories
Home > Documents > 1 Terrorists Face recognition of suspicious and (in most cases) evil homo-sapiens.

1 Terrorists Face recognition of suspicious and (in most cases) evil homo-sapiens.

Date post: 05-Jan-2016
Category:
Upload: harvey-greene
View: 214 times
Download: 1 times
Share this document with a friend
Popular Tags:
35
1 Terrorists Face recognition of suspicious and (in most cases) evil homo-sapiens
Transcript
Page 1: 1 Terrorists Face recognition of suspicious and (in most cases) evil homo-sapiens.

1

Terrorists

Face recognition of suspicious and (in most cases) evil

homo-sapiens

Page 2: 1 Terrorists Face recognition of suspicious and (in most cases) evil homo-sapiens.

2 / 35

The problem

Terrorists need to be identified when passing a security screen

Aim is positive identification of a few faces

Problem is that terrorists try to disguise themselves

Page 3: 1 Terrorists Face recognition of suspicious and (in most cases) evil homo-sapiens.

3 / 35

About Us

Team members: Gülsah Tümüklü (manager) Réka Juhász Emil Szimjanovszki Gergely Windisch

Page 4: 1 Terrorists Face recognition of suspicious and (in most cases) evil homo-sapiens.

4 / 35

Goal

Finding the terrorists Identifying faces even if they are

disguised

Page 5: 1 Terrorists Face recognition of suspicious and (in most cases) evil homo-sapiens.

5 / 35

System

Pre-processing(normalization,

rotation)

Training (eigenface)

Testing (Nearest

Neighbor)

Image Database

Test ImagePre-processing(normalization,

rotation)

Result

Page 6: 1 Terrorists Face recognition of suspicious and (in most cases) evil homo-sapiens.

6 / 35

Our Implementation

Programmed in Matlab Input: RGB image Pre-processing output: Greyscale

image Output: Yes/No (Terrorist-wise)

Page 7: 1 Terrorists Face recognition of suspicious and (in most cases) evil homo-sapiens.

7 / 35

Things We Do

Acquire an Image

Page 8: 1 Terrorists Face recognition of suspicious and (in most cases) evil homo-sapiens.

8 / 35

Things We Do (2)

Locate eyes

Page 9: 1 Terrorists Face recognition of suspicious and (in most cases) evil homo-sapiens.

9 / 35

Things We Do (3)

Normalise (rotate, scale, clip, put eyes to their place) 128*128

Page 10: 1 Terrorists Face recognition of suspicious and (in most cases) evil homo-sapiens.

10 / 35

Things We Do (4)

Face Recognition (details later)

Page 11: 1 Terrorists Face recognition of suspicious and (in most cases) evil homo-sapiens.

11 / 35

Things We Do (n)

Decide, then Call the police

Page 12: 1 Terrorists Face recognition of suspicious and (in most cases) evil homo-sapiens.

12 / 35

101 Useful Tips for Terrorists

Page 13: 1 Terrorists Face recognition of suspicious and (in most cases) evil homo-sapiens.

13 / 35

101 Useful Tips for Terrorists

Page 14: 1 Terrorists Face recognition of suspicious and (in most cases) evil homo-sapiens.

14 / 35

101 Useful Tips for Terrorists

Page 15: 1 Terrorists Face recognition of suspicious and (in most cases) evil homo-sapiens.

15 / 35

101 Useful Tips for Terrorists

Page 16: 1 Terrorists Face recognition of suspicious and (in most cases) evil homo-sapiens.

16 / 35

101 Useful Tips for Terrorists

Page 17: 1 Terrorists Face recognition of suspicious and (in most cases) evil homo-sapiens.

17 / 35

Recognition Part

Problem: Face Recognition Literature about Face Recognition Problems in Face Recognition Eigenfaces

Page 18: 1 Terrorists Face recognition of suspicious and (in most cases) evil homo-sapiens.

18 / 35

Problem: Face Recognition Identifying persons using some priori

information Many potential applications, such as

person identification, human-computer interaction, security systems, image retriveal systems, and finding terrorists

Stages of Face Recognition face detection feature extraction facial image classification

Page 19: 1 Terrorists Face recognition of suspicious and (in most cases) evil homo-sapiens.

19 / 35

Literature about Face Recognition

Classification of Face Recognition Methods: Hollistic Methods Feature-Based Methods Hybrid Methods

Page 20: 1 Terrorists Face recognition of suspicious and (in most cases) evil homo-sapiens.

20 / 35

Face Recognition Methods

Hollistic Methods PCA

Eigenfaces, Probabilistic eigenfaces, Fisherfaces/subspace LDA , SVM, Evolution pursuit, Feature lines, ICA

Other Representations LDA/FLD, PDBNN

Page 21: 1 Terrorists Face recognition of suspicious and (in most cases) evil homo-sapiens.

21 / 35

Face Recognition Methods(cont)

Feature-Based Methods Pure geometry methods Dynamic link architecture Hidden Markov model Convolution Neural Network

Page 22: 1 Terrorists Face recognition of suspicious and (in most cases) evil homo-sapiens.

22 / 35

Face Recognition Methods (cont.)

Hybrid Methods Modular eigenfaces Hybrid LFA Shape-normalized Component-based

Page 23: 1 Terrorists Face recognition of suspicious and (in most cases) evil homo-sapiens.

23 / 35

Problems in Face Recognition Feature Extraction

Global Features Local Features

Handling some problems: Illumination differences Facial expressions Occlusions pose

Page 24: 1 Terrorists Face recognition of suspicious and (in most cases) evil homo-sapiens.

24 / 35

Eigenfaces Firstly introduced by Pentland, and Turk in

1991 It is considered the first working facial

recognition technology Based on PCA Decompose face images into a small set of

characteristic feature images called eigenfaces

Eigenfaces may be thought of as the principal components of the original images

Page 25: 1 Terrorists Face recognition of suspicious and (in most cases) evil homo-sapiens.

25 / 35

Eigenfaces (cont.)

Trainning Part : calculate the Eigenfaces of datases

Classification part : Reconstruct the test image and classify it

Page 26: 1 Terrorists Face recognition of suspicious and (in most cases) evil homo-sapiens.

26 / 35

Calculation of Eigenfaces Calculate average face : v.

Collect difference between training images and average face in matrix A (M by N), where M is the number of pixels and N is the number of images.

The eigenvectors of covariance matrix C (M by M) give the eigenfaces. M is usually big, so this process would be time consuming.

TAAC

],...,,...,,...,[ 111

1 vuvuvuvuA pn

pn

Page 27: 1 Terrorists Face recognition of suspicious and (in most cases) evil homo-sapiens.

27 / 35

Calculation of Eigenfaces(cont.)

Use SVD Substract mean image from training images

diff.images=trainingimages-mean image Find the svd of diff.images [U S V] = svd(diff.images) The columns of U are automatically the e-vectors

of diff.images * diff.images’ Square of S gives eigenvalues

Page 28: 1 Terrorists Face recognition of suspicious and (in most cases) evil homo-sapiens.

28 / 35

Page 29: 1 Terrorists Face recognition of suspicious and (in most cases) evil homo-sapiens.

29 / 35

Classifying a test Image Find the reconstructed image

Calculate weights First find difference test image Diff.test=test Image-mean Image Do inner product of each eigenimage with the

difference image to get a weight vector Find the reconstructed image for m=1:numTrainingImages reconstructionImage = reconstructionImage+(weight(m)*Eimage(:,:,m));

end

Page 30: 1 Terrorists Face recognition of suspicious and (in most cases) evil homo-sapiens.

30 / 35

Classifying a test Image (cont.)

If one of weighs is above a threshold, take the largest one and return that its owner also owns the new face.

Use nearest neighbor method Find minumum distance between

reconstructed image and eigenfaces and assign test image to class which has min distance

Page 31: 1 Terrorists Face recognition of suspicious and (in most cases) evil homo-sapiens.

31 / 35

Pros and Cons

Pros It is fast Efficiency Provides accurate recognition rates

Cons Very sensitive to occlusions,

illuminations, facial expression, pose Only works good with frontal faces

Page 32: 1 Terrorists Face recognition of suspicious and (in most cases) evil homo-sapiens.

32 / 35

Results (1) – Training set

Class 1:(Terrorists)

Class 2:

Class 3:

Page 33: 1 Terrorists Face recognition of suspicious and (in most cases) evil homo-sapiens.

33 / 35

Results (2)

Page 34: 1 Terrorists Face recognition of suspicious and (in most cases) evil homo-sapiens.

34 / 35

Conclusion

Face recognition is a difficult problem

Pre-processing is very important It is not enough to use only global

features Better results can be obtained with

different classifications (eigenfeatures)

Page 35: 1 Terrorists Face recognition of suspicious and (in most cases) evil homo-sapiens.

35 / 35

References M. Turk and A. Pentland. Eigenfaces for recognition. Journal of

Cognitive Neuroscience, 3 (1), 1991a. M. A. Turk and A. P. Pentland. Face recognition using

eigenfaces. In Proc. of Computer Vision and Pattern Recognition, pages 586-591. IEEE, June 1991b.

W. Zhao, R. Chellappa, P. J. Phillips and A. Rosenfeld, "Face Recognition :  A Literature Survey", ACM Computing Surveys(CSUR), vol. 35, issue 4, pp. 399-458, December 2003.

http://cilek.ceng.metu.edu.tr/facedetect B. Galamb: Color Based Eye Location


Recommended