Image authentication by using visual

cryptography

Shabanam Bano Ansari
[email protected]


Abstract: This paper reviews and applies visual cryptography, a
perfectly secure method of keeping images secret, for possible
use in biometric identification and protection. In Biometric
physiological or behavioral characteristics of a person is taken in
image form to authenticate identity of that person.As biometric
template are stored in the centralized database, due to security
threats biometric template may be modified by attacker. If
biometric template is altered authorized user will not be allowed to
access the resource. Visual cryptography is a technique to protect
sensitive image information. It is a secret sharing scheme where a
secret image is encrypted into the n number of shadow images
which independently disclose no information about the original
secret image. And for decryption we have to superimpose these
shadow images and human eye is sufficient to identify secret
image from k number of superimposed shadow images. so visual
cryptography provides security as well as authentication. In this
work, we review some visual cryptography schemes and apply
them to biometric data for the purpose of user authentication

Keywords: Biometrics, Visual cryptography, Authentication,
Shadow image, secret sharing.

Introduction: Increasing demand of Internet and information
resources has a great impact in our everyday life and is making
humans more dependent on computer systems and networks. This
dependency has brought many threats to information security. As
a result, reliably secure mechanisms are required to protect
computers and important information against vulnerabilities like
ID spoofing and unauthorized access to computer resources.
Biometric authentication systems are an example of technologies
which are widely accepted in various applications such as in
identification based systems and access control, ID cards, banking,
etc. Therefore its accuracy and confidentiality should be
guaranteed There are various application where
personal identification is required such as passport control,
computer login control, secure electronic banking, bank ATM,
credit cards, premises access control, border crossing, airport ,
mobile phones, health and social services, etc. Many biometric
techniques are available such as keystroke, odor, ear, hand
geometry, fingerprint, face, retina, iris, palm print, voice and
signature.




Cryptography is a mathematical technique that is used to encrypt
sensitive data, information or message in a form that can be
inferred only by authorized user. Any unauthorized or malicious
user who captures encrypted data or message will not be able to
infer meaning of that data or message. traditional cryptography
like AES, DES or RSA needs well defined encryption and
decryption algorithm and it requires lot of mathematical
calculation so its time consuming. Due to heavy demands of
information in image form, security of image is necessary in
network communication. Visual cryptography is used to address
the problem of image security which is needed in biometric
authentication, banking application. Visual cryptography was
introduced by Naor and Shamir in 1994.It is used to encrypt
written material (printed text, handwritten notes, pictures, etc) in a
perfectly secure way. The decoding is done by the human visual
system directly.Cryptography and data hiding are the well-known
methods for protecting data information by writing it in
unreadable, secret codes and transmitting in a secure way.
Steganography, visual cryptography and watermarking are some
data hiding methods that can be used in order to increase the level
of information security. This paper provides an overview of visual
cryptography as a technique for achieving data security and
discusses its application in biometric authentication. in Section 2,
visual cryptography is introduced and we review the visual secret
sharing schemes proposed by Naor and Shamir Section 3 presents
some applications of visual cryptography to biometric security.
The experimental results and conclusions are presented in Section
4 and section 5.
II. Visual Cryptography :

The visual cryptography scheme (VCS), introduced by Naor and
Shamir in 1994 is a type of secret sharing scheme which can split
secret information into n shares and recover them by
superimposing the shares. In VCS, the secret to be hidden is a
black and white image and each share is compromised of groups
of m black and white sub pixel used to recover a pixel of the secret
image. It is impossible to get any information about the secret
images from shares individually. The other advantage of VCS is
that, unlike other cryptography techniques, this secret recovery
does not need difficult computations. The secret information can
easily be recovered with enough shares and requires human vision
instead of special software or hardware devices
Naor and Shamir proposed a k out of n scheme and assumed that the
image or message is a collection of binary 1 and 0 displayed as black
and white pixels. According to their algorithm, the secret image is turned
into n shares and the secret is revealed if any k of them are stacked
together. So the image remains hidden if fewer than k shares are stacked
together.
Image contrast and the number of sub pixels of the shares and
recovered image are two main parameters in visual cryptography
schemes. The number of sub pixels represents expansion of the
image and should be as small as possible, while the contrast,
which is a relative difference between the maximum value of
Hamming weight for a black pixel and the minimum value of
Hamming weight for white pixel, needs to be as large as possible
Some researchers have focused on contrast degradation and
introduced methods to improve the contrast of the reconstructed
secret image, while many studies have been done on applying
visual cryptography to support grayscale and natural images with
meaningful shares. This is known as extended visual
cryptography.
A. (2,2) Secret Image Sharing Scheme
The basic idea of visual cryptography can be illustrated with the 2-
out-of-2 scheme. In the 2-out-of-2 scheme, every secret pixel of
the image is converted into two shares and recovered by simply
stacking two shares together. This is equivalent to using the OR
operation between the shares. As illustrated in Table1 ,4 sub pixels
are generated from a pixel of the secret image in a way that 2 sub
pixels are white and 2 pixels are black. The pixel selection is a
random selection from each pattern. For example, when the
corresponding pixel is white, one of the rst six rows of Table 1 is
randomly selected to encode the pixel into 2 shares. It is easy to
see that knowing only one share value does not reveal the other
share and the secret image pixel. However superimposing all the
shares reveals the corresponding binary secret image.














TABLE 1
ILLUSTRATION OF A (2,2) SECRET IMAGE SHARING
SCHEME WITH 4 SUBPIXELS
pixel probabilit
y
Share1 Share2 After stacking




white

1/6

1/6
1/6
1/6


1/6
1/6
pixel probability Share1 Share2 After stacking




Black
1/6



1/6
1/6
1/6
1/6
1/6


B. (n, n) visual cryptography
In this type of visual cryptographic scheme, the system generates n
(n 2) number of shares and all shares are needed to be stacked
together to get back the secret information.
C. (K, n) visual cryptography
In this type of visual cryptographic scheme, the system generates n
(n 2) number of shares and at least any k (2kn) shares are
needed to regenerate the secret information.
III. Application of Visual Cryptography To Biometric
Authentication: The ngerprint is the most common human
biometric characteristic that has been used for personal identica-
tion. Results obtained from comparing different biometric traits
show that: the ngerprint has a high value in factors like
permanence, distinctiveness and performance, and medium value
in universality, collectability and acceptability, while the hand-
written signature has the lowest value in universality,
distinctiveness, permanence and performance. For improving
security, reducing fraud and enhancing user convenience,
biometric systems require the process of enrollment, verication
and identication. In enrollment, the biometric template will be
collected and stored in a database for eligible users. Verication is
the process of conrming the authenticity of a biometric sample.
Finally, identication is the process in which the identity of a
biometric sample in a database is determined. Protecting and
securing biometric templates in the database are of great
importance to prevent systems from being vulnerable to some
attacks. Data hiding techniques, such as visual cryptography can
enhance the security by embedding additional information in
biometric images. For improving the quality, increasing the
contrast and simplicity in matching of the reconstructed image.
Figure 4 depicts the encryption and decryption processes of the
ngerprint image using a (2, 2) VCS based on XOR operation
which results in perfect reconstruction of both black and white
pixels.As an example application combining biometrics and data
hiding consider the application of an ID card containing a
biometric template VCS share. Fingerprint images are considered
as a biometric sample and an entrance security system in a bank is
considered as case study in our approach.
Generally], the ngerprint of each eligible person is collected by
the system administrator and they are given to the visual
cryptography algorithm. Random shares are created from
ngerprint images. One of the shares is stored in the database and
the other share is given to the eligible person in the form of
uniqueID card. For verication, the user should insert the IDcard
into the security system. The corresponding share which is stored
in the database is found and stacked with the other random share
that is embedded in the ID card. After verication, the system
should identify the participant. Hence, the system requests the
participants new ngerprints to compare with the minutiae
extracted from secret ngerprint images obtained from the visual
cryptography algorithm. Authentication is accepted if the
matching process succeeds.
The drawback of this method is the limitation in the number of
biometric samples. The ID card requests one secret share for each
biometric template. However, increasing a users biometric
samples or using different types of biometric samples in a
template can lead to increasing the accuracy and security in an
authentication system. Moreover, it makes biometric systems
spoong more difcult. Further, ngerprint authentication systems
may have thousands of users and it is therefore desirable to
minimize the cost and capacity of storing biometric templates in a
database. Using the multiple secret image sharing algorithm is one
way to improve the system

IV. EXPERIMENTAL RESULT
The result of the proposed experiment is implementedin
MATLAB 7.0 running on Windows xp computer. The size of
each of the secret images is 261 127 pixels. As it is shown in
Figure , the images are encrypted into two shares, each with the
size of 522 127 pixels. For decryption, share A and share B are
stacked together and the secret image restored.





Share1
+

Share2
=



V. CONCLUSION
Using biometric characteristics in cryptography has signicant
advantages over traditional cryptographic methods in the case of
authentication. As an example, biometric characteristics of an
individual are difficult to lose, steal or forge. However, biometric
systems are vulnerable to attacks and break-ins by hackers. To
address this issue, some methods are suggested by researchers to
provide the security, accuracy and integrity of biometric
templates in a biometric authentication system. In the
proposed method, authentication in a security system is
achieved by comparing and matching the participants
fingerprints with secret fingerprint images that are derived
from the visual cryptography algorithm.


REFERENCES
[1] M. Naor, and A. Shamir, (1994) Visual Cryptography, Advances in Cryptography-
Eurocrypt '94, vis Lecture Notes in Computer Science 950, pp. 1-12.
[2] Tai-Wen Yue and, Suchen Chiang (2000) A Neural Network Approach for Visual
Cryptography, IEEE-INNS-ENNS International Joint Conference on Neural Networks,
vol.5, pp 494-499.
[3] D. Jena, and S. K .Jena,(2009) A Novel Visual Cryptography Scheme, The 2009
International Conference on Advanced Computer Control, pp-207-211.
[4] C. Hegde, Manu S, P. D.Shenoy, Venugopal K R and L. M. Patnaik (2008), Secure
Authentication using Image Processing and Visual Cryptography for Banking
Applications 16th International Conference on Advanced Computing and
Communication (ADCOM 2008), MIT Campus, Anna University, Chennai, India, pp.
433-439.
[5] A. Adhikari and B. Roy (2008) On some Constructions of Monochrome Visual
Cryptographic Schemes Proceedings of the 1st International Conference on
Information Technology, Gdansk, Poland.
[6] J. K. Pal, J. K. Mandal and K. Dasgupta (2010) A Novel Visual Cryptographic
Technique through Grey Level Inversion (VCTGLI) Proceedings of The Second
International conference on Networks & Communications, Chennai, India, pp. 124-133
[7] M. Heidarinejad, A. A. Yazdi,; K.N. Plataniotis, (2008) Algebraic Visual
Cryptography Scheme for Color Images IEEE International Conference on Acoustics,
Speech and Signal Processing, pp. 1761 1764.
[8] G.R Zhi Zhou Arce,. G. Di Crescenzo (2006) Halftone Visual Cryptography. IEEE
Transactions on Image Processing. , Volume: 15, Issue: 8pp-2441-2453.