Generalized random grid based threshold

visual cryptography with meaningful shares

Devdutt Gautam

Abstract - The meaningful shares are desired in based on encrypting the secret image into ‘n’ shadow
visual cryptography (VC) because they look images and someone having all the ‘n’ images can
natural with increased efficiency and low threat decrypt the message. This scheme was known as ‘k’
of inkling of secret encryptions. The meaningful out of ‘k’ visual secret sharing.
shares are also helpful in dealing with the
management problem of the conventional VC. After further research on visual cryptography, it was
The traditional user-friendly RG-based VCs extended to k out of n visual secret sharing [3]. In
gives advantage of no pixel expansion and no this type of secret sharing someone having any k (or
need for codebook design in generating more) shares out of n shares can decrypt the message
meaningful shares. The current user-friendly but is not visible if less than k shares are stacked
RG-based VCs did not support the general (k, n) together. The original encryption problem is known
threshold VC and in these complementary shares as 2 out of 2 visual secret sharing. The encryption is
may be required. In this paper, a generalized RG- done with the help of the various algorithms and the
based VC with meaningful shares is proposed decryption is done with human visual system in the
and it is compared with the existing meaningful visual cryptography.
share generation algorithm. The proposed The secret images is composed of black and white
scheme inherits the good features of RG-based pixels. The black pixels are denoted by ‘1’ which
existing algorithms. The new feature of the represents opaqueness and white pixels are denoted
proposed scheme is that it supports the (k, n) by ‘0’ which represents darkness.
threshold and provide adaptive visual quality.
Shamir proposed a (k, n) threshold visual secret
sharing in 1979 [5] by encrypting the secret into a
I. Introduction constant coefficient of a random (k-1) degree
Visual Cryptography is the cryptographic technique polynomial. According to Shamir’s scheme the
which is used for encrypting the secret image into secret image can be reconstructed perfectly by using
shadow images so that no one, other than the desired LaGrange’s interpolation. Shamir’s polynomial-
recipient can understand the image information. The based scheme has the advantage that the secret can
shadow images are also called as shares, we can get be recovered lossless. This scheme requires only k
the input image after stacking these shares. There are shares for the reconstruction of distortion-less secret
two types of shares in VC i.e., meaningful shares and image.
non-meaningful shares.
Naor and Shamir [9] first introduced the visual
As the internet is growing very rapidly the digital cryptography in the year 1984.VC is the secret
images can be easily transmitted, obtained, and sharing scheme that allows the decryption without
manufactured by the various tools available on the any computation. In general (k, n) threshold VC
internet. Security of the images becomes a major scheme the secret image is divided into n shares and
concern as the sensitive information is hidden inside we get the decrypted by stacking any k or more than
these images. We can protect the sensitive k shares while any k-1 or less shares did not give any
information from the malicious behaviour in clue about the secret image [9]. There are various
transmission by securing these images. The security applications [12] of VC such as information hiding,
of these images can be ensured with cryptography. watermarking [13], data appending, cyber-crime,
Visual Cryptography is defined as the method of authentication, and identification and transmitting
encoding a secret image into ‘n’ shadow images so passwords.
that the information hidden inside the images cannot The next we will review the random grid-based VC
be used by some unauthorized user. The first visual and after that we introduce our proposed scheme.
cryptography technique was introduced by Moni
Naor and Adi Shamir in 1994.Their scheme was
After Naor and Shamir had introduced the VC, many it can be used in other applications apart from
researchers started finding the problems associated VCs.In this paper we have proposed a meaningful
with VC such as contrast, different formats, and RG-based VC scheme which supports (k, n)
pixel expansion. Random grid (RG)-based VC may threshold VC and it also supports adaptive visual
become the alternative method to overcome the quality. The proposed scheme is compared with one
drawbacks of halftone visual cryptography (HVC) the threshold meaningful RG –based VC given by
Yan et.al.
and extended visual cryptography (EVC). RG-based
VC is used as it has no pixel expansion and no
codebook design is required in this scheme. AFRI
and Keren [1] first proposed the RG-based VC.
According to their scheme the secret image is
encrypted into two meaningless RGs. Kafri and
Karen proposed three algorithms for encrypting the
image into RGs.A secret pixel is taken from the
secret binary image and is encrypted into the
subpixel in each of the two RGs.The sub pixels are
selected randomly form the two columns under the
certain secret pixel. There is the random selection so
that each column is selected with 50% probabilities.
Figure 2- RG based (2,2) scheme application
In that fashion the first subpixel is assigned to RG1
and the followed-up pixel is assigned to RG2 this
process follows up till the last pixel of the secret
image. So, it confirmed that individual share cannot
give any clue about the secret image. When we stack
these shares the black pixels will cover the white
pixels, and the white pixel will decode into the black II. Literature Review
or white pixel with 50% probabilities. This scheme In Visual Cryptography given the image or message,
is known as RG-based (2,2) VC. The size of the n transparencies are generated so that the original
shares is same as that of the secret image. There is image (message) is visible if any k of them are
some loss of contrast in the revealed image but we stacked together. The image remains hidden if fewer
can clearly identify the image. than k transparencies are stacked together. Each
pixel appears within n modified versions (known as
shares) per transparency. The shares are a collection
of m black and white subpixels arranged closely
together. The structure can be described as an n × m
Boolean matrix S. The structure of S can be
described thus: S = (sij)m × n where sij = 1 or 0 i.f.f.
the jth subpixel of the ith share is black or white. The
important parameters of the scheme are: m: the
number of pixels in a share. This represents the loss
in resolution from the original image to the
recovered one. α: the relative difference in the
weight between the combined shares that come from
Figure 1- RG based (2,2) VC secret pixel encryption a white and black pixel in the original image, i.e., the
loss in contrast. γ: the size of the collection of C0
and C1. C0 refers to the sub-pixel patterns in the
Previous RG-based VC does not support meaningful shares for a white pixel and C1 refers to the sub-
shares. Chen and Tsao [17] proposed a friendly RG- pixel patterns in the shares for a black pixel.
based (2,2) VC to exploit meaningful shares in VC.
Chen and Tsao ‘s scheme use the complementary
shares to achieve adjustable visual quality. This
method is not applicable for (k, n) threshold where 1.Visual Cryptography and its Types
k<n. a. Traditional Visual Cryptography
The main motivation of this paper is to propose a Secret sharing [1] is a method by which a secret can
threshold RG-based meaningful VC in which there be distributed between a group of participants,
is no pixel expansion and no codebook design so that whereby each participant is allocated a piece of the
secret. This piece of the secret is known as a share. cryptography schemes suffer mainly because of
The secret can only be reconstructed when enough pixel expansion (any pixel in the original image is
shares are combined. While these shares are expanded during the encryption process to an m×m
separate, no information about the secret can be matrix of pixels) and loss of contrast. A simple
accessed. That is, the shares are completely useless example of (2,3)-EVCS is presented. In this, they
while they are separated. Within a secret sharing refer to a corresponding VCS [1] of an EVCS [8],
scheme, the secret is divided into several shares and we mean a traditional VCS that have the same access
distributed among n persons [2][3]. When any k or structure with the EVCS. Generally, an EVCS takes
more of these persons (where k ≤ n) bring their a secret image and original share images as inputs,
shares together, the secret can be recovered. and outputs shares that satisfy the following three
However, if k - 1 persons attempt to reconstruct the conditions:
secret, they will fail. Due to this threshold scheme,
we typically refer to such a secret sharing system as 1) Any qualified subset of shares can recover the
a (k, n)-threshold scheme or k-out-of-n secret secret image;
sharing. 2) Any forbidden subset of shares cannot obtain any
information of the secret image other than the size
of the secret image;

3) All the shares are meaningful images.

Figure 3- Traditional Visual Cryptography

b. Extended Visual Cryptography

An extended visual cryptography [8] was proposed Figure 4- Extended Visual Cryptography
by Ateniese et al. to protect a binary secret image
There have been many EVCSs proposed in the
with meaningful shares. This is implemented by
literature. Droste, Ateniese et al., and Wang et al.
concatenating an extended matrix to each basis
proposed three EVCSs, respectively, by
matrix [3][4]. The minimum size of the extended
manipulating the share matrices. Nakajima et al.
matrix was obtained from a hypergraph coloring
proposed a (2,2)-EVCS for natural images. Tsai et
model and the scheme was designed for binary
al. proposed a simple EVCS, where its shares were
images only. Extended visual cryptography schemes
simply generated by replacing the white and black
let you construct visual secret sharing schemes in
sub pixels in a traditional VCS share with
which the shared images are meaningful. Image
transparent pixels and pixels from the cover images,
processing means intensive calculations,
respectively. Furthermore, Zhou et al. presented an
reading/writing image files in different formats,
EVCS by using halftoning techniques, [6] and hence
Boolean transformations at the pixel level, and many
can treat Gray-scale input share images. Their
other operations. Visual cryptography [1] represents
methods made use of the complementary images to
a technique for information concealment with major
cover the visual information of the share images.
advantages-it is very low-tech because all that you
Recently, Wang et al. proposed three EVCSs by
need to use it is a couple of transparencies and a
using an error diffusion halftoning technique to
printer. Additionally, the decryption process is
obtain nice looking shares. Their first EVCS also
immediate and does not require any cryptographic
made use of complementary shares to cover the
knowledge from users. To get the secret information,
visual information of the shares. Their second EVCS
participants of the scheme need only properly stack
imported auxiliary black pixels to cover the visual
the transparencies. This presents a high level of
information of the shares. In such a way, each
security (some schemes are absolutely secure),
qualified participants did not necessarily require a
proven in the theoretical studies. That said, visual
pair of complementary share images. Their third
EVCS modified the halftoned share images and stacking blocks to represent the corresponding pixel
imported extra black pixels to cover the visual color in the original image, the problem of circular
information of the shares. However, the limitations permutations occurs. Since two circular
of these EVCSs mentioned above are obvious. The permutations of a stacking block are not considered
first limitation is that the pixel expansion is large. different, two average colors with different
For example, the pixel expansion of the EVCS in is, permutations will be the same in the stacking block
where is the pixel expansion of the secret image and if they have the same combination. Hence the
is the chromatic number of a hyper-graph; in any number of possible color variation is fewer than the
case, the value of satisfies. The construction of the authors claimed 242. Recently, Chang et al.
pixel expansion, where is the number of elements of proposed a color image sharing technique.
which contains exactly elements, and is the set of the
qualified subsets. The second limitation is the bad
visual quality of both the shares and the recovered
secret images. Unfortunately, the EVCS has other
limitations: first it is computation expensive;
second, the void and cluster algorithm make the
positions of the secret pixels dependent on the
content of the share images and hence decrease the
visual quality of the recovered secret image. Third
and most importantly, a pair of complementary
images are required for each qualified subset and the
participants are required to take more than one
shares for some access structures. The same
problems also exist in the first method proposed by
Wang et al. in which each qualified subset does not
require complementary images anymore; however,
this method is only for threshold access structure, Figure 5- Color Mixture In Subtractive Model
and the auxiliary black pixels of their EVCS also
darkened shares.

c. Coloured Visual Cryptography

Rijmen and Preneel had proposed a visual
cryptography approach for color images.[5][9][12]
In their approach, each pixel of the color secret
image is expanded into a 2×2 block to form two
sharing images. Each 2×2 block on the sharing
image is Elledwithred, green, blue and white
(transparent), respectively, and hence no clue about
the secret image can be identified from any one of
these two shares alone. Rijman and Preneel claimed
that there would be 24 possible combinations
according to the permutation of the four colors.
Because human eyes cannot detect the color of a
very tiny subpixel, the four-pixel colors will be
treated as an average color. When stacking the
corresponding blocks of the two shares, there would
be 242 variations of the resultant color for forming a
color image. The approach of Rijmen and Preneel
indeed can produce visual cryptography for color
images. But from the viewpoint of either the additive
model or the subtractive model of chromatology, it
is not appropriate to Ell the blocks with red, green, Table 1- Absorption of Light
blue, and white (transparent) colors. Besides, if we
use the average of the four-pixel colors in the
The algorithm Erst create a palette of a secret image We can decompose each color pixel p into three
and assigns a unique code to each color on the monochromatic components/pixels, namely pc, pm
palette. It then selects two colored cover images, O1 and py, in terms of the three primary colors c, m and
and O2, with size the same as the secret image. y, respectively. In most modern personal computer
Every pixel in the two cover images will be systems, Px is represented as a continuous tone x-
expanded into a block with M =(k × k) subpixels, of colored Gray-level image in which each pixel px is
which M /2 +1 subpixels are randomly selected and colored by one of the 256 shades of color x ranging
Elled with the color of the expanded pixel and the from 0 (white) to 255 (color x), It is easy to
rest are Elled with white (transparent) color. The transform Px into the corresponding x-colored
selection condition is that N positions of the two halftone image Px by applying the halftone
expanded blocks are overlapped, where N is the technology such that each pixel px in Px is either 0
index of the palette of the secret image and is used or color x Hou et al. proposed a method to improve
to indicate the pixel color shared by the two the above drawback. They used the binary encoding
expanded blocks. When recovering the secret image, to represent the sub pixels selected for each block
the algorithm computes the number of the and applied the AND/OR operation randomly to
overlapping sub pixels of every k × k block in the compute the binary code for the stacking sub pixels
two camouflage images and then retrieves the Nth of every block in the cover images. The code ranges
color from the palette to reconstruct the color of the from 0 to 255, but it can be even larger depending
corresponding pixel of the secret image. But this on the expanding factor. Consequently, a secret
method can only deal with a color image with image can be a 256 color or true-color one. Although
limited different colors. For example, if k equals 3, Chang and Hou et al. achieved a certain degree of
M /2 +1 is at most 5, which is obviously too small sharing color image information, the drawback is
and unreasonably restrictive for today’s that secret images must be decrypted with heavy
applications. computation, which would violate the principle of
visual cryptography that uses human eyes to decrypt
secret images. Hou et al. used the concepts of color
decomposition and contrast adjustment to produce
two shares needed by visual cryptography.
Overlapping these two shares will reveal the secret
information automatically. Although this method
requires no mass computation to reconstruct secret
images, it is nonetheless difficult to obtain totally
random noise shares. Some image boundaries might
be found on each share, thus compromising the
secrecy required. In this method, they combined the
previous results in visual cryptography, the halftone
technology, and the color decomposition principle to
develop algorithms of visual cryptography for gray-
level and color images. The method retains the
advantage of traditional visual cryptography,
namely, decrypting secret images by human eyes
without any cryptography computation. For
information security, it also ensures that hackers
cannot perceive any clue about the secret image
from any individual sharing image.

Table 2- Possible colors mixed by (C,Y) and M-colored

halftone pixels
 were achieved by using better error diffusion
techniques [7], the technique proposed in satisfies
the following 3 requirements: (i) a secret image
should be a natural image, (ii) images that carry a
secret image should be a high-quality natural image
and (iii) computational cost should be low. This
technique is based which satisfies both (ii) and (iii)
and to satisfy (i), introduces an additional feedback
mechanism into the secret image embedding process
to improve the quality of the visually decoded secret
image. Methods described in only satisfy part of the
three requirements. The area of contrast within
Figure 6- Colored Visual Cryptography
halftone and grayscale VC is an interesting one
because the contrast determines exactly how clear
the recovered visual secret is. Cimato et al.
d. Halftone Visual Cryptography developed a visual cryptography scheme with ideal
This method of secret sharing expands on Naor and contrast by using a technique known as reversing,
Shamir's original endings in the 2-out-of2 secret which was originally discussed by. Reversing
sharing scheme[5]. It also takes extended visual changes black pixels to white pixels and viceversa.
cryptography a step further. The halftoning Viet and Kurosawa's scheme allows for perfect
technique that is used can be applied to color and restoration of the black pixels but only almost
grayscale images. Halftoning simulates a continuous perfect restoration of the white pixels. Cimato et al.
tone through the use of dots, varying either in size or provide their results for perfect restoration of both
in spacing. Based on the idea of extended visual black and white pixels. Each share also contained a
cryptography, Zhou et al. [8] set about improving smaller amount of information than Viet and
these techniques by proposing halftone grayscale Kurosawa's which makes it a more desirable and
images which carry significant visual information. secure scheme. Yang et al. [12] also looked at
Traditional VC [1] produces random patterns of dots reversing and the shortcomings of Viet and
with no visual meaning until the shares are Kurosawa's scheme. Their work presented a scheme
superimposed. This raises the suspicion of data that allowed perfect contrast reconstruction based on
encryption. Halftoning attempts to alleviate this any traditional visual cryptography sharing scheme.
suspicion by having visually pleasing attributes.
This means creating halftone shares that carry one
piece of information, such as another image, while
having the secret hidden until both shares are
superimposed. This gives no indication that any
encryption has been performed on both shares. This
drastically improves the security model for visual
cryptography. Along with Zhou, present novel
techniques by which halftone images can be shared
with significant visual meaning which have a higher
quality than those presented within by employing
error diffusion technique [7]. These error diffusion
techniques spread the pixels as homogeneously as
possible to achieve the improvements in the shares
overall quality. A halftone scheme was proposed in
which the quality of the shares is improved by using
contrast enhancement techniques. However, the
problem with this scheme is that it is not perfectly
secure. By using a space celling curve ordered
dithering technique, grayscale images can be
converted into an approximate binary image. This Figure 7- Halftone Visual Cryptography
allows encryption and decryption of the Gray-level
images using traditional visual cryptography
methods. Further improvements made in this area
e. Random Grid based Visual random grids R1 and R2 which reveal B when
Cryptography superimposed where Rk[i, j] ∈ {0, 1} (transparent or
Random grid (RG)-based VC [2][3][4] maybe an opaque), 1_i_w,1_j _h and k ∈ {1, 2}Encryption(B)
alternative method to overcome the drawbacks,
since RG-based VC has no pixel expansion and
requires no codebook design. RG-based VC was
first presented by Kafri and Keren, which encrypts
the secret image into two meaningless RGs. To
illustrate the principles of RG-based VC, one of the
three distinct encryption algorithms presented by
Kafri and Keren [2] is shown. Each secret pixel
taken from a secret binary image is encrypted into
one sub pixel in each of the two RGs. The sub pixels
are randomly selected from the two columns
tabulated under the certain secret pixel. The
selection is random so that each column is selected
with the same probabilities (50%). Then, the first
sub pixel is assigned to RG1 and the following sub
pixel is assigned to RG2. Thus, an individual share
gives no clue about the secret image. When the sub
pixels are stacked, the opaque (black) pixels will
cover the transparent (white) pixels. The black secret
pixel will be decoded into black pixel, and the white
secret pixel will be decoded into white pixel or black
pixel with the same probabilities (50%). As a result,
the secret could be revealed by HVS. The secret is
encrypted into two random shares which have the
same size as the secret image. The revealed image is Table 3- Probability of occurrence of pixels
clearly identified, although some contrast loss
occurs. Follow-up investigations on RG-based VC
[3] were discussed to extend the features of RG-
Depending on the further three algorithms for the
based VC, such as contrast, color images, threshold,
generation of the share images which is a part of
(n, n) threshold and (k, n) threshold. Unfortunately,
encryption based on probabilistic approach.
the previous RG-based VC does not support
meaningful shares. To exploit meaningful shares in
RG-based VC, Chen and Tsao proposed a friendly
RG-based (2, 2) VC by designing a procedure of Algorithm 1
distinguishing different light transmissions on the 1. Generate R1 as a random grid, T(R1) = ½
two shares. However, in Chen and Tsao's user- // for (each pixel R1[i, j], 1_i_w and 1_j _h) do
friendly RG-based (2, 2) VC [4], complementary // R1[i, j] = random_pixel (0, 1)
shares are used to achieve adjustable visual quality. 2. for (each pixel B[i, j ], 1_i_w and 1_j _h) do 2.1 {
In addition, the method is not for (k, n) threshold, if (B[i, j] = 0) R2[i, j] = R1[i, j] else R2[i, j] = R1[i,
where k<n. Kafri and Keren proposed three different j] }
algorithms to accomplish the encryption for the 3. output (R1,R2)
binary images. Given a secret binary image B, these
algorithms as follows produce two random grids R1
and R2 such that they leak no information of B Algorithm 2
individually, yet they reveal B in our visual system 1. Generate R1 as a random grid, T(R1) = 1/2
when superimposed. 2. for (each pixel B[i, j ], 1_i_w and 1_j _h) do
2.1 { if (B[i, j] = 0) R2[i, j] = R1[i, j ] else R2[i, j] =
random pixel(0, 1) }
3. output (R1,R2)
Algorithm
Encryption of a binary image by random grids Input:
A w ×h binary image B where B[i, j] ∈ {0, 1}(white
or black), 1_i_w and 1_j _h Output: Two shares of
Algorithm 3 the image is visible but it is very blurry. As more of
1. Generate R1 as a random grid, T(R1) = 1/2 the image is downloaded, the clearer the resulting
2. for (each pixel B [i, j], 1_i_w and 1_j _h) do image becomes, until it is fully loaded. The
2.1 {if (B [i, j] = 0) R2[i, j] = random pixel (0, 1) annoying presence of the loss of contrast makes
else R2[i, j] = R1[i, j] } traditional visual cryptography schemes practical
3. output (R1,R2) only when quality is not an issue which is relatively
rare. Therefore, the basic scheme is extended to
Note that random pixel (0, 1) is a function that allow visual cryptography to be directly applied on
returns a binary value 0 or 1 to represent a grayscale and color images. Image halftoning [5] is
transparent or opaque pixel, respectively, by a coin- employed to transform the original image from the
flip procedure and R1[i, j] denotes the inverse of grayscale or color space into the monochrome space
R1[i, j]. The three algorithms are capsulated into a which has proved to be quite effective. To further
generic procedure named Encryption so that when improve the quality, artifacts introduced in the
Encryption (B) is called, each of Algorithms 1, 2 or3 process of halftoning have been reduced by inverse
can be applied onto B. Also note that in this, 0 (1) halftoning. With the use of halftoning and a novel
denotes a white (black) pixel in the secret binary micro block encoding scheme, the technique has a
image or a transparent (opaque) pixel in the unique flexibility that enables a single encryption of
encrypted share interchangeably. a color image but enables three types of decryptions
on the same ciphertext. The three different types of
decryptions enable the recovery of the image of
varying qualities. The physical transparency
stacking type of decryption enables the recovery of
the traditional VC quality image. An enhanced
stacking technique enables the decryption into a
halftone quality image. A progressive mechanism is
established to share color images at multiple
resolutions. Shares are extracted from each
resolution layer to construct a hierarchical structure;
the images of different resolutions can then be
restored by stacking the different shared images
together.
Figure 8- Random grid based Visual Cryptography

f. Progressive Visual Cryptography

Progressive VC [13] takes into consideration the
premise of perfect secret recovery and high-quality
secret reconstruction. Many of the schemes do
require computational effort to perfectly reconstruct
the secret. A technique that enables visual
cryptography to be used on color and grayscale
images is developed in progressive color visual
cryptography. Many current states of the art visual
cryptography techniques lead to the degradation in
Figure 9- Progressive Visual Cryptography
the quality of the decoded images, which makes it
unsuitable for digital media (image, video) sharing
and protection. In this, a series of visual
cryptography schemes have been proposed which 2. Application of Visual Cryptography
not only support grayscale and color images, but Finding practical and useful applications of this
also allow high quality images including that of technology has proved to be somewhat of an issue
perfect (original) quality to be reconstructed. The over the past fifteen years. In this we attempt to
meaning of the progressive term refers to how the provide some possible applications for VC in
final image is built up. For example, when commercial technologies, specifically within the
downloading or viewing an image on a web page, secure printing industry.
the image is loaded in stages. The full dimension of
a. Morie Patterns Visual cryptography has been implemented using
A potential application for visual cryptography is its Moire patterns. Desmedt and Le provide a scheme
use in conjunction with Moire patterns. Moire by which secrecy and anonymity are both satisfied.
patterns (or fringes) are induced when a revealing Moire patterns occur when high frequency lattices
layer such as a dot screen or line grating is are combined to produce low frequency lattice
superimposed on top of a periodically repeating patterns. It is the difference in these high frequencies
shape. The resulting Moire pattern is changing any that give the Moire patterns. Figure9 shows an
of the following geometric parameters example of these Moire patterns.
characterizing the individual grid structures, namely
period, orientation, and shape. Whether a dot screen
or a line grating is used, both induce Moire fringes
with the same geometric properties. The revealing
layer contains horizontal black lines (line grating),
between those lines is transparent white space.
When the revealing layer is superimposed, the
shapes that appear are the magnified versions of the
repeating pattern. Figure8 demonstrates this
magnifying effect. This magnifying property could
be used as a method of locating hidden VC [2][3] Figure 12- Superimposing of Moiré Patterns
shares within a Moire pattern. This magnification
factor of these patterns can be calculated as follows,
let p represent the period of shapes in the base layer,
the period of the line gratings in the revealing layer The Moire cryptography model is as follows: The
is denoted as pr. For the magnification to work, the embedded (secret) image is randomized into two
periods must be sufficiently close. When the shares, known as pre-shares. These are independent
revealing layer is superimposed, the repeating of the original image. XORing [9] these pre-shares
pattern in the base layer is stretched along the will recover the original. Next, the hiding algorithm
vertical axis. There is no change in the horizontal takes the cover image and combines it with each of
axis. This magnification can be represented as pm. the pre-shares separately. Its output is the final two
The following equation expresses this magnification shares that are used to reveal the original embedded
along the vertical axis: image. These resulting shares look the same as the
input cover image that is used. There are three
different Moire schemes proposed by Desmedt and
Le, lattice rotation, lattice smooth rotation, and dot
orientation. The problem with lattice rotation is that
the boundary between differently-rotated areas in
the shares becomes visible. However, this scheme
produced very sharp decrypted ciphertext. Lattice
smooth rotation fixed the boundary issues but
introduced another problem, namely, the artifacts
introduced into the shares stand out too much and
become visible. The pair settled on the final scheme,
Figure 10- Moire Pattern dot orientation, as their chosen implementation. The
dots from the shares are converted into diamond
shape \dots”, this makes for a less visible boundary
than circular or elliptical dots. The scheme encodes
a white pixel by superimposing two squares onto the
shares whose dots are oriented at different angles. To
encode a black pixel, dot patterns are used that are
of the same angle. This produces two different Moire
patterns for the white and black dots. That means
this scheme uses the Moire patterns to recover the
secret embedded image, rather than traditional
visual cryptography schemes which use the gray
level of the squares to recover the secret.
Figure 11- Magnified Repeating Pattern
b. Watermarking watermarking techniques. Although watermarking is
Sonny and Philip invented the “copyright series designed to protect copyright by embedding secret
management device” to protect the copyright of to the host media, according to the differences of
digital audio cassettes in 1980’s, it is recognized as functionalities and appearance of watermarking,
the first device to protect copyright of digital watermarking has been grouped into many
commercial products. The aim of this product is to categories.
protect the ownership of users and encourage them
to create the products of themselves. In 1996, Adobe
Systems Inc. added watermarking functionalities in
Adobe Photoshop 4.0 which was developed by III. Methodology
Digimarc Inc. At the same time, the institutes of
NEC completed the software: Tiger Mark Data
Blade, Informix Software finished watermarking Algorithm 1:
functionalities in the database product Encryption of a binary image by random grid.
InformixUniversal Server (Information
Input: A w × h binary image S where S [i, j] ∈ {0,
Management System). Europe electronic industries
1} (white or black), 1<=i <=w and 1<=j <=h.
hope to monitor illegally copy video and audio
commercial products using watermark censoring Output: Two shares of random grids SC1 and SC2
system to and the illegal duplications. The project is which reveal S when superimposed. Also, the shares
called TALISMAN (Tracing Authors’ Rights by were of the same size as that of the actual binary
Labelling Image Services and Monitoring Access image.
Networks), it started from September 1995, 11
communication and broadcasting companies, 1.Generate a random grid SC1 with the transparency
research institutes and universities involve in it. of 0.5 i.e. (T (R1) = ½). For (each pixel SC1 [i, j],
There are two important ID adopted in the system: 1<=i <=w and 1<=j <=h)
ID of the copyright owner which is embedded in the
2. For each position in the binary image ( B[i, j ],
multimedia data, another is unique international
1<=i <=w and 1<=j <=h) do { if (B[i, j ] = 0) SC2[i,
code such as ISBN. The two IDs are expected to
j ] = SC1[i, j ] else SC2[i, j ] = SC1[i, j ] }
work jointly so as to protect the ownership. Since
1996, the International Workshop of Information 3. Output (SC1, SC2)
Hiding was held every year averagely, watermarking
research has become a hot research area in the past
ten years, watermarking has been a very important
Output: Using OR Gate
issue. In the past ten years, in the conferences of
ACM, IEEE, and IFIP, watermarking is the main
research topics for media security and assurance. In
1995, Cox et al. extended watermarking algorithm
from spatial domain, and presented a spread
spectrum watermarking. Even today, this paper has
been widely cited. It is thought of as the important
milestone in watermarking research, it’s the
landmark of robust watermarking. Audio and Video
watermarking are also very important members in
the family. In 2000, Horvatic et al. presents work
about audio watermarking, this approach combined
scrambling and DCT transformation together, it
guarantees the quality of host audio while ensure the
requirements of robustness. The research scientists Figure 13- Shares Generated by Algorithm 1
from Microsoft research Asian applied
watermarking in video and presented the video
watermarking based on wavelets. In ACM
SIGGRAPH ‘99, Praun et al. presented robust
watermarking for meshes, and introduced
watermarking to computer graphics. Before this
work, Benedens, Ohuchi, Yeung et. al. also studied
this issue; Y. J. Song et al. compare the usual
 Step4: Select the B[i] where I belong to
{0,1,2,3,4……n} randomly either 1 or 2.
Step5: Now randomly arrange b1, b2, b3…to the
shares.
Step6: Output the n shares SC1, SC2, SC3…...SCn.

Figure 14- Recovered Secret Algorithm 1

Figure 15- Shares Generated by Algorithm 4

Algorithm 2:
Threshold random grid based visual secret
sharing system.
Basically, in this algorithm we are trying to do three
things at once and this is an advance and more
complex algorithm.

1.Generation of n different shares for an individual

image and the value of n is selected by the user itself.

2.Implementing the threshold parameter on the

stacking procedure of the image such that below the Figure 16- Recovered secret by Algorithm 4 using OR
threshold value no image is generated or no stacking
information is produce.

3.Resolving the participant increasing issue in the

VSS. Output: Using AND Gate

Algorithm

Input: A w×h binary image S where S[i, j ]∈{0, 1}

;1<=i <=w and 1<=j <=h .
Output: n non-meaningful shares SC1, SC2, SC3,
……. SCn.

Step1: For each position (i, j) where 1<=i<=w and

1<=j <=h. Repeat step 2-5.
Step2: Generate an array of b1, b2, b3……bk.
Where b[i] belongs to {0,1} repeat step 3 for w
belongs to {1,2,3,4…………nk}.
Step3: If share SC[i] not belongs to B[i] then
randomly select p and flip b. Figure 17- Shares generated by algorithm 4
 Figure 18- Recovered Secret by Algorithm 4 using AND
stacking

IV. Implementation
Algorithm:
Input: A binary secret image S and a cover image C
both with M*N pixels, the threshold parameters
(k,n) and light transmission parameters w0 ,w1
,where 0<w0≤w1<1.

Output: n meaningful shares SC1, SC2…SCn.

STEP1: For each position (i, j) € {(i, j) | 1 ≤ i ≤ M,1

≤ j ≤ N} in the secret image, repeat steps 2-6.

STEP2: Set 𝑏̃1= S (i, j), repeat step 3 for k-2 times,
i.e. for p=1, 2, k-2, to generate pixels b1, b2…….
bk-2, 𝑏k-1; where bx and 𝑏̃x denote the temporary
pixels x=1, 2…., n-1, n.

STEP3: If 𝑏̃p=0, 𝑏̃p+1= bp; otherwise 𝑏̃p+1=𝑏̅p.

Where bp is generated randomly by using flipcoin
function. Figure 19- Stacking of 3 shares

STEP4: If the corresponding cover image pixel

C(i,j)=0 ,w=w0 otherwise w=w1.

STEP5: Generate a pixel bk-1 by bk-1=g(w).If 𝑏̃k-

1=0 then, bq=bk-1 , q=k,k+1,…….,n ;
otherwise=g(w) ,q=k,k+1,……..,n

STEP6: The order of n pixels b1, b2…..., bn-1, bn

are rearranged and the rearranged n pixels are
assigned to SC1(i, j), SC2(i, j) ………SCn (i, j).

STEP7: Output the n shadow images SC1,

SC2……..., SCn.

Figure 20- Stacking of 4 shares

Output: Meaningful Shares
 out that secret is transmitted apart from the receiver
and the sender.

VI. References

[1] Naor, M., & Shamir, A. (1994, May). Visual

cryptography. In Workshop on the Theory
and Application of Cryptographic Techniques (pp.
1-12). Springer, Berlin, Heidelberg.
[2] Kafri, O., & Keren, E. (1987). Encryption of
Figure 21- Stacking of 5 shares
pictures and shapes by random grids. Optics
letters, 12(6), 377-379.
[3] Shyong Jian Shyu. (2006),” Image encryption by
random grids”, Ming Chuan University,5
V. Conclusion Der Ming Rd. Gawi Shan, Taoyan,333.
This paper proposes a Generalized random grid- [4] Shyu, S. J. (2007). Image encryption by random
based threshold visual cryptography with grids. Pattern Recognition, 40(3), 1014-
meaningful shares. In this technique we have taken 1031.
a secret image and converted it into two shares. The [5] Chen, T. H., Tsao, K. H., & Wei, K. C. (2008,
image is taken as input i.e., it may be a colored November). Multiple-image encryption by
image or a gray image. It is then converted into a rotating random grids. In Intelligent Systems Design
binary image. We generate two shares the first is and Applications, 2008. ISDA'08. Eighth
randomly generated and the second one is generated International Conference on (Vol. 3, pp. 252-256).
by applying some conditions on binary image and IEEE.
the first share. We obtain the image when we stack [6] T-H Chen K-H Tsao, Image Encryption by (n,n)
both the shares. The stacking of the shares is done random grids ,in :Proceedings of
by using OR gate. A share on an individual does not Eighteenth Information Security Conference, IEEE,
convey any information about the original image. Haulien 2008.
We have used the concept of progressive visual [7] Chen, T. H., & Tsao, K. H. (2009). Visual secret
cryptography means that as we impose more and sharing by random grids revisited.
more shares on each other we can get the better Pattern recognition, 42(9), 2203-2217.
visual quality of the recovered secret image. [8] Shyu, S. J. (2009). Image encryption by multiple
Random grids are the matrices of the black and white random grids. Pattern Recognition,
pixels which can simply decrypt the secret image by 42(7), 1582-1596.
using the stack operation and no computation is [9] El-Latif, A. A. A., Yan, X., Li, L., Wang, N.,
required by the computer. We computed the contrast Peng, J. L., & Niu, X. (2013). A new
of the image by applying the process on it. To meaningful secret sharing scheme based on random
improve the visual quality of the image we have used grids, error diffusion and chaotic
modified gates. This paper possesses the complete encryption. Optics & Laser Technology, 54, 389-
knowledge about visual cryptography, it is types, 400.
applications. It comprises of all the necessary [10] X. Wu. W. Sun, Generalized random grids and
algorithms used for visual secret sharing using its applications in visual
generalized random grids. Apart from that cryptographies Trans. Inf. Forensics Secur (2013)
meaningful share generation algorithm [16] is also [11] Teng Guoa, b,*,Feng Liua
proposed. By using the meaningful shares, we can , Chuankun Wua
increase the efficiency and security of the images. ,(2013)”Threshold visual secret sharing by
The shares generated in meaningful shares are like random grids with improved contrast”.
natural shares. We have used two images for [12] K out of k Extended Visual Cryptography
meaningful shares one is secret image and the other Scheme by Random Grids: Teng Guo,
is cover image. The secret image is the secret which Feng Liu, Chaunkun Wu
we want to transmit and the cover image is used for [13] Ou, D., Sun, W., & Wu, X. (2015). Non-
hiding the secret image so that any one cannot find expansible XOR-based visual cryptography
scheme with meaningful shares. Signal Processing,
108, 604-621.
[14] Yan, X., Wang, S., Niu, X., & Yang, C. N.
(2015). Random grid-based visual secret
sharing with multiple decryptions. Journal of Visual
Communication and Image
Representation, 26, 94-104.
[15] Chiu, P. L., & Lee, K. H. (2015). User-friendly
threshold visual cryptography with
complementary cover images. Signal processing,
108, 476-488.
[16] Yan, X., Wang, S., Niu, X., & Yang, C. N.
(2015). Generalized random grids-based
threshold visual cryptography with meaningful
shares. Signal Processing, 109, 317-333.
[17] Hou, Y. C., Quan, Z. Y., & Tsai, C. F. (2015). A
privilege-based visual secret sharing
model. Journal of Visual Communication and Image
Representation, 33, 358-367.
[18] Xuehu Yan, Yuliang Lu. (2016). Participants
increasing for threshold random
grid—based visual secret sharing.
[19] Chao, H. C., & Fan, T. Y. (2017). XOR-based
progressive visual secret sharing using
generalized random grids. Displays,49, 6-15.
[20] Yan, X., Liu, X., & Yang, C. N. (2018). An
enhanced threshold visual secret sharing
based on random grids. Journal of real-time image
processing, 14(1), 61-73.
[21] Anjney Pandey, Subhranil Som, “Applications
and Usage of Visual Cryptography: A
Review”.
[22] Jiun-Tsung Wang, Wen-Hsiang Tsai,” Data
Hiding in PDF files and Applications By
Imperceivable Modifications of PDF Object
Parameters”.
[23] Jonathan Weir, WeiQi Yan,” Visual
Cryptography and Its Applications”
[24] T-H Chen K-H Tsao, Threshold visual secret
sharing by random grids,
J. Syst.Softw.84(7) (2011).
[25] T.H Chen, K.H Tsao, User-Friendly random
grid based visual secret sharing. IEEE
Trans. Circuit Syst. Video Tech. 21 (11) (2011).