Seeing and Believing is a Threat: A Visual Cryptography Schemes

INTRODUCTION

Pace of development of network technology has helped the man in sending and receiving any digital data over the Internet and on the GO. The Information security places a very vital role in providing the safer path for the data between two users without the third’s intervention and a one of the ways is by image secret sharing. Even though Cryptography (Schneier, 2007) and image steganography is the good candidates, where, the former needs special decryption device and the latter never reveals its presence.

In the conventional cryptography techniques, the problems being the image been seen by the intruders is solved by many effective algorithms. However, still the decryption of the encrypted cipher text need strict computations making it a costly and lesser effective of the hardware/software for the decoding has direct proportions to the security algorithm, hence making the relation of security and cost/efficiency directly proportional.

To tackle the above problems, Noar and Shamir (1995) found visual cryptography, which deals the humans vision ability in the decryption process (Wu and Chen, 1998). The secret is differentiated (broken into parts called shares) and when seen individually reveals nothing (Lukac and Plataniotis, 2005) but when the transparencies, i.e., shares are integrated; the original secret message is visible to human visual system.

The Visual cryptography which merely makes use of human insight is much useful in scenarios where neither computational devices, nor cryptographic knowledge is not available or usage is impossible. This particular attribute of Visual Cryptography, rules over all other cryptographic techniques for sharing secret images securely.

There are many works available in the literature for binary image (Noar and Shamir, 1995) secret sharing, gray image (Lin and Tsai, 2003) as secret or even to colour images (Hou, 2003; Leung et al., 2009). Visual cryptography is available for different purposes like secret image sharing (Shyu et al., 2007), authentication (Chen et al., 2012), visual digital signature scheme (Jaafar and Samsudin, 2013), Cheat immune and traceable visual cryptography scheme (Yuan et al., 2012) and copyright protection (Chen et al., 2009) like watermarking (Jin and Kim, 2012).

After carefully reviewing the available literature, this study has been suggested to know about Visual cryptography and its variants for sharing single secret to multiple users and multiple secret to multiple users in binary and gray level images.

LITERATURE REVIEW

(k, n)-Threshold scheme: For the n receivers the secret image is divided to n shares by the famous k,n-threshold procedure. The shares contain only an nth portion of the complete image and when combined, the complete picture i.e., secret image appears. A simple schematic diagram has been illustrated in Fig. 1, its the method from Lukac and Plataniotis (2005), Firstly, by the process of half toning (dithering), the, input image gets converted as binary form having spatial coordinates x and y whose range is from 1, 2, up to K1 and 1, 2, up to K2, respectively. Then the image is shared. The original pixel r(x, y) (1 for white, 0 for black) is expanded to a set of smaller pixels (m1xm2) through a function for encryption to yield shares of the n receivers. As the spatial arrangement deviates for each block useful data is not unveiled until and unless an access to the predefined number of shares has been obtained.

A reference pixel r (i, j) sited at (i, j) (of original image) is mapped into blocks of size m1xm2 by an encrypting function, say, fe(.). For illustration, let us consider a block of 2x2 for 2, 2 routine:

The encryption procedure is as follows:

The sets:

and

Fig. 1:	Visual cryptography schema

Fig. 2(a-f):	A sample result-modified (k, n) threshold scheme for 4 users, (a) Input, (b) Share 1, (c)Share 2, (d) Share 3, (e) Share 4 and (f) Output

The entire matrices from the column permutation of basis matrices nxm₁m₂ are taken into account:

Some sample results are presented in Fig. 2 for 4 such users.

Spreading factor m1, m2 decides the matrices’ dimension in addition to n (nxm₁m₂) gives the participants’ number. How much a share is big when compared to the (original) image is also based on this spreading factor. Hence, m1, m2 must be as small as possible for optimum functioning. If the secret pixel is white it may belong to set C1 and else if it is black then the S1, S2 must be chosen from set C0. Random number generator is employed for the choice. Intended for this 2, 2 plots, all pixels in both S1 and S2 are tantamount if secret pixel is white and S1 should be complement of that in S2 if the pixel is black.

Decryption is done by a simple process of stacking these shares. The original secret image can be viewed by human visual ability, without the need of any complex computations. This study reveals four steps: half toning, encrypting and decrypting.

Advantages:

Drawback: