Unauthorized Access and Key Generation from Face Feature Data

Download Full-Text PDF Cite this Publication

Text Only Version

Unauthorized Access and Key Generation from Face Feature Data

Dr. Kamini Solanki1, Abhishek Mehta2

1Associate Professor (Parul Institute of Computer Application, Parul University, India)

2Assistant Professor (Parul Institute of Computer Application, Parul University, India)

AbstractFace Recognition is used for security purpose in real time application. This paper represents two modules. In first module, face recognition is done by combining local binary pattern (LBP) and principal component analysis (PCA) in different way. Proposed algorithm is used for better recognition rate. PCA reduce dimension of image and LBP describe the local texture of image data. So proposed method will increase the recognition rate (RR) of face and also decreased false match rate (FMR) but there is no difference in verification time and also fastest key generation algorithm from face feature vectors. So it is most suitable for real time application. We compared proposed method with both PCA and LBP to figure out changes. In case of execution time, there is no difference between of existing and proposed method. In Second module, key generation from feature vectors of proposed algorithm for the purpose of the security. According to the result proposed algorithm generate fastest key compared to the existing key.

Index Terms: Face Recognition¸ principal component analysis, Unauthorized Access, Key Generation

INTRODUCTION

Facial features are done by global features and local features of face. Global features focus on whole entire facial image so it has less accuracy. Local features focus only on local area of the face. So it is most accurate method. The local binary pattern (LBP) is design for local texture description and shape of an image. It is done by dividing an image into small parts from which local features are extracted. These local features consist of binary patterns which describe the location of pixels in those areas.

The gained features from the regions are combined into a single feature histogram, which is represents the image. Images can be compared by distance measurement.

It is based on local feature extraction and dimension reduction. In proposed method face recognition done by combining the LBP and PCA in different way which provides very good results. It increased the performance of recognition. It can work against face images with different facial conditions.

Fig.1 Face recognition methods

face recognition problem using PCA Algorithms

Illumination Problem

Illumination problem happens when same image with condition. So person have to keep with fix lighting condition, fixed distance, same facial expression and also same view point. It can emerge extensively different when lighting condition is different. [8]

Pose Problem

Face recognition with different facial poses that is called pose problem. If face rotation made very large changes in face appearance it reduce recognition rate. If person try to match same image with different facial pose, it show the different result.[8]

face recognition problem using LBP Operator

Limitation of the basic LBP operator is that its small 3 * 3 neighborhood cannot capture dominant features with large- scale structures. It cannot deal with the texture at different scales and the operator was later generalized to use neighborhoods of different size so LBP cannot work well on large scale images.[5]

database

The FEI face database is a Brazilian face database that holds 14 images for each of 200 individuals, a total of 2800 images. All images are colorful and standing frontal position with profile rotation up to about 180 degrees. Age of persons between 19 to 40 years old with distinct appearance, hairstyle, and adorns.[47]

object of proposed method and local binary pattern Increased recognition rate

Decreased false mate rate

Biometric template security using cryptography Fastest key generation from facial feature

Local Binary Pattern (LBP) operator describes the local texture of gray scale image. LBP is a binary code for an image-pixel which describe local neighborhood of that pixel. This operator works with the eight neighbors of a pixel, using the value of this center pixel as a threshold. If a neighbor pixel has a higher gray value than the center pixel (or the same gray value) than a one is assigned to that pixel, else it gets a zero. The LBP code for the center pixel is then produced by concatenating the eight ones or zeros to a binary code.[43]

Fig. 2 The original LBP operator

Fig. 3 LBP on FEI Database

PCA ALGORITHM

Principal Component Analysis (PCA) is well-organized method for face recognition. It is most popular methods for a face image. It is used to reduce the dimension of the image and also holds variations in the image data. It projecting face image into a feature space that covers the variations significantly. Those features are called as Eigen faces, because they are the eigenvectors or Principal Component of faces.

PCA has several steps [8][20][19]

Stage 1: Subtract the Mean of the data from each variable (our adjusted data) subtraction of the overall mean from each of our values as for covariance we need at least two dimensions of data. It is in fact the subtraction of the mean of each row from each element in that row.

Stage 2: Calculate and form a covariance Matrix

Stage 3: Calculate Eigenvectors and Eigen values from the covariance Matrix Eigen values are a product of multiplying matrices however they are as special case. Eigen values are found by

multiples of the covariance matrix by a vector in two dimensional space (i.e. a Eigenvector). This makes the covariance matrix the equivalent of a transformation matrix. Stage 4: Chose a Feature Vector (a fancy name for a matrix of vectors) Once Eigenvectors are found from the covariance matrix, the next step is to order them by Eigen value, highest to lowest. This gives you the components in order of significance. Here the data can be compressed and the weaker vectors are removed producing a lossy compression method, the data lost is deemed to be insignificant.

Stage 5: Multiply the transposed Feature Vectors by the transposed adjusted data. The final stage in PCA is to take the

transpose of the feature vector matrix and multiply it with the transposed adjusted data set (the adjusted data set is from Stage 1 where the mean was subtracted from the data).

The technique proposed in this paper was tested, verified and implemented with C# framework for number of systems. The software purchased cannot be installed on client system without the verification and validation of the watermarked information. If anybody wants to pirate the copy of software of the client on its system, the proposed technique does not allow him/her to do so, if implemented. This has given an opportunity to client to purchase the software and use it without the risk of redeployment of software to others. By doing so, intellectual property of the developer and value for money of the client, both are protected. This technique can be implemented over different mobile platforms.

PCA on FEI Database

Fig. 4 PCA on FEI Database

PROPOSED ALGORITHM

Step 1: Find the mean image.

Step 2: Reallocate the array for the local binary pattern. Step 3: LBP feature extraction using Feature Vectors. Step 4: Calculate the eigenvectors and eigenvalues

Step 5: Combine LBP local feature vectors with PCA global feature vectors.

Step 6: Key generation using these face features and encrypt image data using Euclidean and Euclidean Squard Distance Metrics method.

Step 7: Encrypted image data stored in database. Step 8: Apply Step 1 to 6 if it is Inputted Image.

Step 9: Decrypt Database all image data using same key.

Step 10: Verification of Database images and inputted image using Ecudian distance measurement method.

Step 11: Retrieved image from database which have minimum distance between input image and Database images.

IMOLIMENTATION

Implementation of proposed algorithm

PC A

LB P

PC A

LB P

PCA

LBP

05

5

6

76.

66

52.

22

23.

34

47.

78

0.22

0.04

10

5

6

63.

33

48.

66

36.

67

51.

34

0.25

0.05

15

5

6

58.

88

46.

66

41.

12

53.

34

0.27

0.05

20

5

6

56.

66

46.

55

43.

34

53.

45

0.30

0.06

PC A

LB P

PC A

LB P

PCA

LBP

05

5

6

76.

66

52.

22

23.

34

47.

78

0.22

0.04

10

5

6

63.

33

48.

66

36.

67

51.

34

0.25

0.05

15

5

6

58.

88

46.

66

41.

12

53.

34

0.27

0.05

20

5

6

56.

66

46.

55

43.

34

53.

45

0.30

0.06

Fig.5 Implementation of proposed algorithm

Key Generation and Implementation of image cryptography

Fig. 6 Implementation of proposed algorithm on real time database

RESULT AND ANALYSIS OF PCA AND LBP ALGORITHM

PCA & LBP on FEI Face Database

Subje ct

DB

Image s

Traine e Image s

RR%

FMR%

Avg.Verificati on Time

(In Seconds)

PCA

LBP

PCA

LBP

PCA

LBP

05

14

14

98.5

0

62.8

5

1.5

37.1

5

0.37

0.05

10

14

14

95.7

1

70.0

0

4.29

30.0

0

0.25

0.05

15

14

14

92.8

5

79.5

2

7.15

20.4

8

0.29

0.05

20

14

14

89.6

4

64.6

4

10.3

6

35.3

6

0.32

0.05

Subj ect

DB

Ima ges

Inpu t Ima ges

RR%

FMR%

Avg.Verific ation Time (In Seconds)

Subj ect

DB

Ima ges

Inpu t Ima ges

RR%

FMR%

Avg.Verific ation Time (In Seconds)

Table 1. PCA & LBP on FEI Face Database PCA& LBP on Real time Database (6 trainee images)

Table 2. PCA& LBP on Real time Database (6 trainee images)

PCA& LBP on Real time Database (14 trainee images)

Subj ect

DB

Ima ges

Inpu t Ima ges

RR%

FMR%

Avg.Verific ation Time (In Seconds)

PC A

LB P

PC A

LB P

PCA

LBP

05

5

14

82.

85

68.

00

17.

15

32.

00

0.04

0.04

10

5

14

75.

71

65.

00

24.

29

35.

00

0.05

0.05

15

5

14

76.

66

70.

47

23.

34

29.

53

0.05

0.06

20

5

14

75.

71

69.

64

24.

29

30.

36

0.06

0.06

Table 3. PCA& LBP on Real time Database (14 trainee images)

proposed system on real time database (14 trainee images)

Subje ct

Databa se Images

Input Imag es

Recogniti on Rate (RR in

%)

Fals e Matc h Rate (FM

R in

%)

Avg.Verificat ion Time

(In Seconds)

05

5

14

92.85

7.15

0.04

10

5

14

90.71

9.29

0.06

15

5

14

84.28

15.7

2

0.04

20

5

14

84.64

15.3

6

0.07

Table 4 proposed system on real time database (14 trainee images)

Subje ct

Databa se Images

Input Imag es

Recogniti on Rate (RR in

%)

Fals e Matc h Rate (FM

R in

Avg.Verificat ion Time

(In Seconds)

Subje ct

Databa se Images

Input Imag es

Recogniti on Rate (RR in

%)

Fals e Matc h Rate (FM

R in

Avg.Verificat ion Time

(In Seconds)

proposed system on real time database (3 DB images)

Table 5 proposed system on real time database (3 DB images)

Comparison (Difference) of proposed algorithm with PCA and LBP

Table 6 Comparison (Difference) of proposed algorithm with PCA and LBP

Key Generation, Encryption, and Decryption of existing algorithm

Subject

DB

Images

Input Images

Key Generation

Encryption

Decryption

(In seconds)

05

14

14

0.05

0.15

0.14

10

14

14

0.05

0.14

0.14

15

14

14

0.06

0.14

0.14

20

14

14

0.08

0.22

0.15

Table 7 Key Generation, Encryption, and Decryption of existing algorithm

Key Generation, Encryption, and Decryption of proposed algorithm

Subject

DB

Images

Input Images

Key Generation

Encryption

Decryption

(In seconds)

05

14

14

0.00

0.12

0.12

10

14

14

0.00

0.12

0.12

15

14

14

0.00

0.13

0.11

20

14

14

0.00

0.13

0.11

Table 8 Key Generation, Encryption, and Decryption of proposed algorithm

Difference between Key Generation, Encryption, and Decryption of existing algorithm and proposed algorithm

Subject

DB

Images

Input Images

Key Generation

Encryption

Decryption

(Decreased In seconds)

05

14

14

0.05

0.03

0.02

10

14

14

0.05

0.02

0.02

15

14

14

0.06

0.01

0.03

20

14

14

0.08

0.09

0.04

Table 9 Difference between Key Generation, Encryption, and Decryption of existing algorithm and proposed algorithm

CONCLUSION

%)

05

3

14

88.57

11.4

3

0.03

10

3

14

79.28

20.7

1

0.04

15

3

14

71.90

28.1

0

0.05

20

3

14

72.85

27.1

5

0.05

%)

05

3

14

88.57

11.4

3

0.03

10

3

14

79.28

20.7

1

0.04

15

3

14

71.90

28.1

0

0.05

20

3

14

72.85

27.1

5

0.05

LBP is fastest execution operator so it is most suitable for real time application and remove illumination problem but it works only on local regional of image. So it cannot detain main features of large-scale structures. PCA has high accuracy rate but it has illumination problem and pose problem. If LBP feature vector combine with PCA Eigen vector according to this way, it remove the illumination and pose problem and also increased recognition rate and decreased false match rate as well as not much more difference between verification time. For the security purpose, image data is encrypted with the proposed feature vectors so it is used for the security of database template.

Subje ct

DB

Image s

Input Image s

Increased RR%

Decreased FMR%

Diff. AVG.Veri.Ti me

(In Seconds)

PCA

LBP

PCA

LBP

PCA

LBP

05

5

14

10.0

0

24.8

5

10.0

0

24.8

5

00.00

00.00

10

5

14

15.0

0

25.7

1

15.0

0

25.7

1

00.01

00.01

15

5

14

07.6

2

13.8

1

07.6

2

13.8

1

00.01

00.02

20

5

14

08.9

3

15.0

0

08.9

3

15.0

0

00.01

00.01

Subje ct

DB

Image s

Input Image s

Increased RR%

Decreased FMR%

Diff. AVG.Veri.Ti me

(In Seconds)

PCA

LBP

PCA

LBP

PCA

LBP

05

5

14

10.0

0

24.8

5

10.0

0

24.8

5

00.00

00.00

10

5

14

15.0

0

25.7

1

15.0

0

25.7

1

00.01

00.01

15

5

14

07.6

2

13.8

1

07.6

2

13.8

1

00.01

00.02

20

5

14

08.9

3

15.0

0

08.9

3

15.0

0

00.01

00.01

REFERENCES

  1. Hardik Kadiya, Comparative Study on Face Recognition Using HGPP, PCA, LDA, ICA and SVM, Global Journal of Computer Science and Technology Graphics & Vision Volume 12 Issue 15 Version 1.0 Year 2012 Type: Double Blind Peer Reviewed International Research Journal Publisher: Global Journals Inc. (USA) Online ISSN: 0975-4172 & Print ISSN: 0975-4350 , Merchant Engineering College.

  2. Vinay Rishiwal, Ashutosh Gupta, Improved PCA Algorithm for Face Recognition, World Applied Programming, Vol (2), Issue (1), January 2012. 55-59 Special section for proceeding of International e-Conference on Computer Engineering (IeCCE) 2012 ISSN: 2222-2510 ©2011 WAP journal. www.waprogramming.com

  3. Prof. B.S PATIL1 Prof. A.R YARDI2, ,DrMrs Patil S B3, REAL TIME FACE RECOGNITION BY VARING NUMBER OF EIGENVALUES, International Journal of Advanced Scientific and Technical Research Issue 3 volume 1, January- February 013 Available online on http://www.rspublication.com/ijst/index.html ISSN 2249-9954

  4. LiorRokach , Chapter 15 CLUSTERING METHODS Department of Industrial Engineering Tel-Aviv University liorr@eng.tau.ac.il Oded Maimon Department of Industrial Engineering Tel-Aviv University maimon@eng.tau.ac.il.

  5. Maneesh Upmanyu, Anoop M. Namboodiri, Kannan Srinathan, and C. V. Jawahar, Blind Authentication: A Secure Crypto- Biometric Verification Protocol, IEEE TRANSACTIONS ON INFORMATION FORENSICS AND SECURITY, VOL. 5, NO. 2, JUNE 2010

  6. Wilman W. W. Zou, Very Low Resolution Face Recognition Problem Student Member, IEEE, and Pong C. Yuen, Senior Member, IEEE, IEEE TRANSACTIONS ON IMAGE PROCESSING, VOL. 21, NO. 1, JANUARY 2012

  7. PAUL VIOLA , MICHAEL J. JONES , Real-Time Face Detection ,International Journal of Computer Vision 57(2), 137 154, 2004 Kluwer Academic Publishers. Manufactured in The Netherlands. Robust viola@microsoft.com, mjones@merl.com

  8. Abhishek Nagar, Student Member, IEEE, Karthik Nandakumar, Member, IEEE, and AnilK , Multibiometric Cryptosystems Based on Feature-Level Fusion. Jain, Fellow, IEEE, IEEE TRANSACTIONS ON INFORMATION FORENSICS AND SECURITY, VOL. 7, NO. 1, FEBRUARY 2012.

  9. Koen Simoens, Julien Bringer, HervéChabanne, and StefaanSeys, A Framework for Analyzing Template Security and Privacy in Biometric Authentication Systems, IEEE TRANSACTIONS ON INFORMATION FORENSICS AND SECURITY, VOL. 7, NO. 2, APRIL 2012

  10. Jianfeng Ren a,n, XudongJiang b, Junsong Yuan b , A complete and fully automated face verification system on mobile devices a BeingThere Centre,InstistuteforMediaInnovation,NanyangTechnologicalUniv ersity,50NanyangDrive,Singapore637553,Singapore b Electrical &ElectronicEngineering,NanyangTechnologicalUniversity, NanyangLink,Singapore639798,Singapore, Elsevier

    ,www.elsevier.com/locate/pr

  11. Thomas Heseltine, Face Recognition: A Literature Review, DPhil Research Student University of York ,2012

  12. AliJaved Faculty of Telecom & Information Engineering, University of Engineering & Technology, Face Recognition Based on Principal Component Analysis Taxila, I.J. Image, Graphics and Signal Processing, 2013, 2, 38-44 Published Online February 2013 in MECS (http://www.mecs-press.org/) DOI: 10.5815/ijigsp.2013.02.06

  13. Anil K. Jain, Brendan Klare and Unsang Park Department of Computer Science and Engineering Michigan State University East Lansing, MI, U.S.A, Face Recognition: Some Challenges in Forensics {jain, klarebre, parkunsa}@cse.msu.edu

  14. Andrew Wagner, Student Member, IEEE, John Wright, Member, IEEE, Arvind Ganesh, Student Member, IEEE, Zihan Zhou, Student Member, IEEE, Hossein Mobahi, and Yi Ma, Senior Member, IEEE, Towards a Practical Face Recognition System: Robust Alignment and Illumination by Sparse Representation

  15. K. Fukunaga (1989) Statistical Pattern Recognition New York:

    Academic Press, 1989

  16. M. Gu, S.C. Eisenstat (1994) A Stable and Fast Algorithm for Updating the Singular Value Decomposition Research Report YALE DCR/RR-996, 1994, Yale University ,New Haven, CT

  17. S. Chandrasekaran, B.S. Manjunath, Y.F.Wang, J. Winkeler, and

    H. Zhang (1997) An Eigenspace update algorithm for image analysis, journal of Graphical Model and Image Processing, 1997

  18. A.L. Yuille, D.S. Cohen, and P.W. Hallinan (19889) Feature extraction from faces using deformable templates proc. CVPR, San Diego, CA, June 1989.

  19. Vinay Rishiwal1 Ashutosh Gupta2, Improved PCA Algorithm for Face Recognition, World Applied Programming, Vol (2), Issue (1), January 2012. 55-59,Special section for proceeding of International e-Conference on Computer Engineering (IeCCE) 2012,ISSN: 2222-2510 ©2011 WAP journal. www.waprogramming.com

  20. Prof. B.S PATIL1 Prof. A.R YARDI2, ,DrMrs Patil S B, REAL TIME FACE RECOGNITION BY VARING NUMBER OF EIGENVALUES ,International Journal of Advanced Scientific and Technical Research Issue 3 volume 1, January-February 2013 Available online on http://www.rspublication.com/ijst/index.html ISSN 2249-9954

Leave a Reply

Your email address will not be published. Required fields are marked *