International Journal of Engineering Trends and Technology (IJETT) – Volume 4 Issue 8- August 2013
Encapsulated Visual Cryptography Strategies for
Data Sharing
G. Manthru Naik#1, Suresh Babu Siriparapu #2
#1
II Year M.Tech Student, #2Assistant Professor
#1, #2
CSE Department, Sri Mittapalli College of Engineering,
Guntur, Andhra Pradesh, India.
Abstract — A (k, n)-threshold visual cryptography strategy is
proposed to encode a secret image into n shadow images, where
any k or more of them can visually recover the secret image, but
any k-1 or fewer of them gain no information about it. The
decoding process of a visual cryptography strategy, which differs
from traditional data sharing, does not need complicated
cryptographic mechanisms and computations. This kind of
strategy is very useful as the participants in such security systems
need not know the cryptographic knowledge in order to recover
the secret image from the shares. This phenomenon is known as
VCS (Visual Cryptography Strategy). An encapsulated VCS is
the one which is capable of generating meaningful shares when
compared with the shares of the VCS. This paper proposes
construction of EVCS by inserting the random shares (result of
VCS) into covering images. The experimental results revealed
that the proposed EVCS is more secure and flexible than the
EVCSs found in literature. VIP synchronization retains the
positions of pixels carrying visual information of original images
throughout the color channels and error diffusion generates
shares pleasant to human eyes. Comparisons with previous
approaches show the superior performance of the new method.
Keywords— EVCS (Encapsulated Visual Cryptography
Strategy), Data Sharing, VIP Synchronization, VCS.
I.
INTRODUCTION
How to keep a secret is always an important issue in many
applications. Two major approaches to this aim are
information hiding and data sharing. A common method for
information hiding is to use the watermarking technique (Cox
et al., 2001; Katzenbeisser and Petitcolas, 2000; Johnson et
al., 2000). And a well-known technique for data sharing is the
cryptography method proposed by Shamir (1979).
Unfortunately; these applications are all restricted to the use of
binary images as input due to the nature of the model. This
drastically decreases the applicability of visual cryptography
because binary images are usually restricted to represent textlike messages. The characteristics of images for showing
object shapes, positions, and brightness thus cannot be
utilized. At the age of Internet, data of image form gradually
replace data of text form. It is not sufficient that visual
cryptography strategies can only deal with binary images. The
underlying operation of this strategy is logical operation OR.
In this paper, we call a VCS with random shares the
traditional VCS or simply the VCS. In general, a traditional
VCS takes a secret image as input, and outputs n shares that
satisfy two conditions: 1) any qualified subset of shares can
ISSN: 2231-5381
recover the secret image; 2) any forbidden subset of shares
cannot obtain any information of the secret image other than
the size of the secret image. An example of traditional (2, 2)VCS [8] can be found in Fig. 1, where, generally speaking, a
(k, n) – VCS means any k out of n shares could recover the
secret image. In the strategy of Fig. 1, shares (a) and (b) are
distributed to two participants secretly, and each participant
cannot get any information about the secret image, but after
stacking shares (a) and (b), the secret image [10] can be
observed visually by the participants. Many other applications
of VCS, other than its original objective (i.e., sharing secret
image), have been found, for example, authentication and
identification [4], watermarking [5] and transmitting
passwords [6] etc.
The term of encapsulated visual cryptography strategy
(EVCS) was first introduced by Naor et al. in [3], where a
simple example of (2, 2)-EVCS was presented. Generally, an
EVCS takes a secret image and n original share images as
inputs, and outputs n shares that satisfy the following three
conditions: 1) any qualified subset of shares can recover the
secret image; 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. EVCS
can also be treated as a technique of steganography. One
scenario of the applications of EVCS is to avoid the custom
inspections, because the shares of EVCS are meaningful
images.
Fig 1: Shares and the recovered secret image of an embedded (3, 3)EVCS after reducing the black ratios.
The pixel expansion of the (k, n) – EVCS in [13] is m+m0
where m0 ≥ [n / (k-1)]. The second limitation is the bad visual
quality of both the shares and the recovered secret images
[11]; this is confirmed by the comparisons in [16].
Unfortunately, the EVCS in [16] has other limitations: first it
http://www.ijettjournal.org
Page 3590
International Journal of Engineering Trends and Technology (IJETT) – Volume 4 Issue 8- August 2013
is computation expensive; second, the void and cluster
algorithm makes 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, which will
inevitably cause the attentions of the watchdogs at the custom
and increase the participants’ burden. The same problems also
exist in the first method proposed by Wang et al. [17]. The
limitation of this method is that the visual effect of each share
will be affected by the content of other shares, and the content
of the input original share images should be chosen in a
selected way.
Fig 6: Proposed (2, 2)-EVCS for fine share images
II. TRADITIONAL VCS & H ALFTONING TECHNIQUE
In this section, we give some definitions about VCS and
some preliminary results about the halftoning technique by
using the dithering matrix. Suppose all the participants of a
secret sharing scheme [1], [2] are V = {0, 1, 2,…….., n-1}. The
qualified and forbidden subsets are represented as (ΓQ, ΓF)
where ΓQ is the superset of qualified subsets, and ΓF is the
superset of forbidden subsets. The minimal qualified access
structure (Γmin) and the maximum forbidden access structure
(ΓMAX) are computed as follows:
(Γmin) =
{A Q  Q}
(ΓMAX) =
{A F  F}
Fig 2: Experimental results of (2, 2)-EVCS proposed in [19], [16].
Fig 3: Experimental results of Method 2 & 3 for (2, 2)-EVCS in [17].
Lastly, the EVCS proposed in [14] is only for (2, 2)
access structure; besides their limitations on the access
structure, the scheme may have security issues when relaxing
the constraint of the dynamic range. (Explicit discussions on
the security of the EVCS in [14] can be found in [18, Sec.
4.2].) This paper proposes an embedded EVCS scheme with
overall good properties. In this, we will show that our scheme
has competitive visual quality compared with many of the
well-known EVCSs. Besides, our EVCS has many specific
advantages against these well-known EVCSs, respectively.
Fig 4: Proposed (2, 2)-EVCS.
Fig 5: Experimental results of the second method proposed in [17] for
fine share images
ISSN: 2231-5381
When the dealer encodes a secret pixel they just needs to
randomly choose a share matrix for a white (respectively,
black) pixel and distributes the row, containing sub pixels, to
the participant. In other words, a share matrix can be seen as a
random column permutation of the basis matrix. When
decoding the secret pixel, the participants only need to stack
their shares, i.e., they print their shares on the transparencies
and stack the transparencies. Then the secret pixel can be
observed visually by human eyes. This approach of
construction of VCS will have small memory requirements (it
keeps only the basis matrices) and it is efficient (to choose a
matrix as it only needs to generate a permutation of the basis
matrices randomly).
III. DITHERING MATRIX FOR HALFTONING
TECHNIQUE
The main drawback of the VCS’s proposed in [3], [8],
[7], and [12] is that, they cannot deal with the gray-scale
image. MacPherson [18] proposed a VCS to deal with the
gray-scale image; however, it has large pixel expansion. In
order to deal with the gray-scale image, the halftoning
technique or dithering technique was introduced into the
visual cryptography [9], [14], [19], [20], [21] which is used to
convert the gray-scale image into the binary image. This
technique has been extensively used in printing applications
which has been proved to be very effective. Many kinds of
halftone algorithms have been proposed in the literature [25].
In this paper, we make use of the patterning dithering [22]
which makes use of a certain percentage of black and white
pixels, often called patterns, to achieve a sense of gray scale in
the overall point of view. The pattern consists of black and
white pixels, where different percentages of the black pixels
stand for the different grayness’s. The halftoning process is to
map the gray-scale pixels from the original image into the
patterns with certain percentage of black pixels.
http://www.ijettjournal.org
Page 3591
International Journal of Engineering Trends and Technology (IJETT) – Volume 4 Issue 8- August 2013
Once meaningful covering shares are created using the
dithering matrices, the understanding of insertion process is
described in algorithm 2.
Fig 7: Halftoned patterns of the dithering matrix of gray levels 0-9
A more efficient way is by using the dithering matrix.
The dithering matrix is an m×n integer matrix, denoted as M.
The entries, denoted as Ma, b for 0 ≤ a ≤ m-1 and 0 ≤ a ≤ n-1,
of the dithering matrix are integers between 0 and mn-1,
which stand for the gray-levels in the dithering matrix. The
halftoning process is formally described in Algorithm 1.
Algorithm 1: The halftoning process for each pixel:
Input: The m×n dithering matrix M and a pixel y with gray
level g in an input image.
Output: The halftoned pattern at the position of the pixel y.
For a = 0 to m-1 do
For b = 0 to n-1 do
If g ≤ Ma, b then print a black pixel at position (a, b);
else
Print a white pixel at position (a, b).
IV. BASIS FOR PROPOSED EVCS
In this section, we will give an overview of our
construction. The main idea in the proposed EVCS is to use the
VCS to encode secret image. First, it ensures that the secret
image can be visually observed by stacking a qualified subset
of shares. And second, it ensures that the shares are all
meaningful in the sense that parts of the information of the
original share images are preserved. The definition is
reasonable, because the information of the original share
images is not as important as that of the secret image for the
participants. The basis of EVCS contains two main steps:
1) Create x covering shares, represented as x0, x1, …, xn-1.
2)Creating encapsulated shares by embedding the
corresponding VCS into the x covering shares, represented
as z0, z1, …, zn-1
CREATION OF COVERING SHARES WITH THE HELP
OF DITHERING MATRICES
Normally, in order to create the covering shares for the
universal access structure, we need to create dithering
matrices. We acquire the covering shares fulfilling that the
stacking consequences of the capable covering shares are all
black images by halftoning the input original share images.
Characterize the points of the dithering matrix as the basics in
the regular set. The regular set contains all the gray-levels in
the dithering matrix, is the halftone pixel extension. We
signify the sets as divisions; each division corresponds to an
accomplice and a covering share.
By using the two concepts, these matrices can be derived:
1) Generating the Covering Subsets with Minimum Average
Black Ratio.
2) Converting the Covering Subsets into Dithering Matrices.
Algorithm 2: Embedding Process:
Input: The covering shares that constructed in the above
section, those corresponding VCS (C0, C1) with
pixel extension and the secret image J.
Output: The n embedded shares z0, z1, z2,…, zn-1.
Step 1: Dividing the covering shares into blocks that
contains the t (≥m) sub pixels each.
Step 2: Choose m embedding positions in each block in the
n covering shares.
Step 3: For each black (respectively, white) pixel in J,
randomly choose a share matrix M  C1.
Step 4: Embed the m sub pixels of each row of the share
matrix M into the m embedding positions chosen in
Step 2.
It is easy to note that the share pixel expansion can be
different from the secret image pixel expansion. The secret
image pixel expansion is independent of the share pixel
expansion. Because we can choose the block size to be
arbitrarily large, the secret image pixel expansion can be
arbitrarily large. In our scheme, because the original share
images are only expanded when they are halftoned, the share
pixel expansion is equal to the halftone pixel expansion. To
avoid the image distortion during the halftoning process, we
usually let be a square number.
V.
VI. EMBEDDING PROCESS OF RELATIVE VCS INTO
COVERING SHARES
ISSN: 2231-5381
Fig 8: Result of Algorithm 2.
As can be seen in the above figure, it is evident that first
of all original VCS encryption is applied on secret image in
order to generate random shares which have no visual
meaning. Then the converting shares are generated by taking
some images as input. Afterwards, the random shares are
embedded into covering shares by following steps given in
algorithm 2. The result of the embedding is the collection of
meaningful covering shares.
http://www.ijettjournal.org
Page 3592
International Journal of Engineering Trends and Technology (IJETT) – Volume 4 Issue 8- August 2013
VII. ADDITIONAL ADVANCEMENTS ON VISUAL
QUALITY OF SHARES
By reducing the black ratio of covering sets in
the images, we can increase the visual quality of the
shares. Recall that in the embedding process, out of
every sub pixels in the covering shares are replaced
by the sub pixels of the basis matrix of the
corresponding VCS. Hence, there is no difference
whether these sub pixels are all black or not in the
stacking result of the qualified covering shares. The
advancements can be done with the help of two
methods i.e.
a) Reduce the Black Ratio of the Covering Subsets
for s = t.
The black ratio requires the gray-levels of all the pixels
in the original input images. So, for an input image, the dealer
needs first to darken the input image to satisfy the
requirement. If the black ratio is high, the darkening process
will decrease the visual quality of the covering shares, so the
black ratio is expected to be as small as possible. Recall that in
the embedding process, out of every sub pixels in the covering
shares are replaced by the sub pixels of the basis matrix of the
corresponding VCS. Hence, there is no difference whether
these sub pixels are all black or not in the stacking result of
the qualified covering shares.
Creation of dithering matrix with reduced black
ratio:
Step 1: Choose the m (<s) embedding positions in the starting
dithering matrix, and denote the gray-levels in the
embedding positions as (g0, g1,…, gm-1). Remove these
positions from the universal set G, and denote the new
universal set as G’, i.e., the rest gray-levels other than
that in the embedding positions.
Step 2: Generate covering subsets Ai ’ for the universal set G’,
by using the methods proposed, where i=0, 1 …n-1.
Step 3: Convert covering subsets Ai’ into dithering matrix Di’
by using the methods proposed, where i=0, 1 …n-1.
Step 4: For each dithering matrix Di’, swap the gray-levels in
the embedding positions with gray-levels in a similar
way and denote the final dithering matrix as Di, where
i=0, 1 …n-1.
b) Reduce the Black Ratio of the Covering Subsets
for s ≠ t.
This method is to construct lcm(s, t)/s dithering matrices
for the input original share image. The dithering matrices are
used to halftone adjacent pixels of the input original share
images at a time. The dithering matrices can be divided into
blocks with sub pixels for each block; we embed secret sub
pixels into each block. Hence each dithering matrix has a
ISSN: 2231-5381
different universal set. For each universal set, we construct the
dithering matrix by using the method that is similar to the
above creation, respectively.
Creation of lcm(s, t)/s dithering matrices for each
input original share image for s ≠ t:
Step 1: Concatenate lcm(s, t)/s starting dithering matrices with
s entries, and divide these starting dithering matrices
into lcm(s, t)/t blocks.
Step 2: Choose the m embedding positions in each block.
Step 3: Concatenate the lcm(s, t)/t blocks, and divide them
into lcm(s, t)/s dithering matrices.
Step 4: For each dithering matrix, remove the embedding
positions, and the rest of the positions in each
dithering matrix constitute the universal set for this
dithering matrix.
Step 5: Generate the dithering matrixes according to the above
creation.
VIII.
EXPERIMENTAL OUTCOMES AND ASSESSMENTS
To measure visual quality of covering shares, we
provide two well-known objective numerical measurements,
the peak signal-to-noise ratio (PSNR) and the universal
quality index (UQI) [23]. In such a way, the PSNR values can
reflect the effects of a combination of the following possible
processes in EVCSs: darkening, embedding, and modification.
The PSNR is defined as follows:
Where MSE is mean square error. The UQI is adopted to
assess the distortion of each share image with its original
gray-scale share image (after being scaled to the size of
shares). Hence, the UQI value can reflect the effect of the
halftoning process besides that of the darkening, embedding
and modification processes in EVCSs [15]. The formal
definition of UQI can be found in [23].
To better show the advantages of the proposed scheme,
we give further comparisons on fine share images. We only
consider EVCSs that the contents of each share are
independent of the others. From this point of view, Wang et
al.’s second EVCS is the most competitive one. Note that, the
share images ruler for the proposed scheme and Wang et al.’s
scheme are both generated from the input image ruler, where
the line width of both share images is 1.
IX. CONCLUSION
In this paper, we proposed a construction of EVCS
which was realized by embedding the random shares into the
meaningful covering shares. An input gray-level image is first
converted into an approximate binary image with the dithering
technique, and a visual cryptography method [24] for binary
images is then applied to the resulting dither image. This
scheme possesses the advantages of inheriting any developed
cryptographic technique for binary images and having less
http://www.ijettjournal.org
Page 3593
International Journal of Engineering Trends and Technology (IJETT) – Volume 4 Issue 8- August 2013
increase of image size in ordinary situations. The decoded
images can reveal most details of original images
Some of the advantages of our scheme include the
ability to work with gray scale images; pixel expansion is
smaller; always secure; no necessity for complementary
shares. General access structure can be applied and each
participant needs only one share to be carried. The proposed
scheme is flexible and there are trade-offs between visual
quality and secret image pixel expansion and between visual
quality and share pixel expansion. The experimental results
revealed that the proposed scheme is capable of providing
high visual quality of shares which is competitive when
compared with other EVCSs reviewed in the literature.
Furthermore, our construction is flexible in the sense
that there exist two trade-offs between the share pixel
expansion and the visual quality of the shares and between the
secret image pixel expansion and the visual quality of the
shares. Comparisons on the experimental results show that the
visual quality of the share of the proposed embedded EVCS is
competitive with that of many of the well-known EVCSs in
the literature.
ACKNOWLEDGEMENT
We, authors express gratitude to all the
anonymous reviewers for their affirmative
annotations among our paper. Thanks to every
reviewer for reviewing our paper and give valuable
suggestions.
REFERENCES
[1] A. Shamir, “How to share a secret,” Commun. ACM, vol.
22, no. 11, pp. 612–613, 1979.
[2] G. R. Blakley, “Safeguarding cryptographic keys,” in
Proc. National Computer Conf., 1979, vol. 48, pp. 313–
317.
[3] M. Naor and A. Shamir, “Visual cryptography,” in Proc.
EUROCRYPT’ 94, Berlin, Germany, 1995, vol. 950, pp.
1–12, Springer-Verlag, LNCS.
[4] M. Naor and B. Pinkas, “Visual authentication and
identification,” in Proc. CRYPTO’97, 1997, vol. 1294, pp.
322–336, Springer-Verlag LNCS.
[5] T. H. Chen and D. S. Tsai, “Owner-customer right
protection mechanism using a watermarking scheme and
a watermarking protocol,” Pattern Recognit., vol. 39, pp.
1530– 541, 2006.
[6] P. Tuyls, T. Kevenaar, G. J. Schrijen, T. Staring, and M.
Van Dijk, “Security displays enable secure
communications,” in Proc. First Int. Conf. Pervasive
Computing, Boppard Germany, Springer-Verlag Berlin
LNCS, 2004, vol. 2802, pp. 271–284.
[7] C. Blundo, A. De Bonis, and A. De Santis, “Improved
schemes for visual cryptography,” Designs, Codes and
Cryptography, vol. 24, pp. 255–278, 2001
[8] N. K. Prakash and S. Govindaraju, “Visual secret sharing
schemes for color images using halftoning,” in Proc. Int.
Conf. Computational Intelligence and Multimedia
Applications (ICCIMA 2007), 2007, vol. 3, pp. 174–178.
ISSN: 2231-5381
[9] Y. C. Hou, “Visual cryptography for color images,”
Pattern Recognit., vol. 1773, pp. 1–11, 2003.
[10] S. J. Shyu, S. Y. Huang, Y. K. Lee,R. Z.Wang,
andK.Chen, “Sharing multiple secrets in visual
cryptography,” Pattern Recognit., vol. 40, no. 12, pp.
3633–3651, 2007.
[11] P. A. Eisen and D. R. Stinson, “Threshold visual
cryptography schemes with specified whiteness levels of
reconstructed
pixels,”
Designs,
Codes
and
Cryptography, vol. 25, pp. 15–61, 2002.
[12] G. Ateniese, C. Blundo, A. De Santis, and D. R. Stinson,
“Extended capabilities for visual cryptography,” ACM
Theoretical Comput. Sci., vol. 250, no. 1–2, pp. 143–
161, 2001.
[13] D. S. Wang, F. Yi, and X. B. Li, “On general
construction for extended visual cryptography schemes,”
Pattern Recognit., vol. 42, pp.3071–3082, 2009.
[14] M. Nakajima and Y. Yamaguchi, “Extended visual
cryptography for natural images,” in Proc. WSCG Conf.
2002, 2002, pp. 303–412.
[15] D. S. Tsai, T. Chenc, and G. Horng, “On generating
meaningful shares in visual secret sharing scheme,”
Imag. Sci. J., vol. 56, pp. 49–55, 2008.
[16] Z. Zhou, G. R. Arce, and G. Di Crescenzo, “Halftone
visual cryptography,”IEEE Trans. Image Process., vol.
15, no. 8, pp. 2441–2453, Aug. 2006.
[17] Z.M.Wang, G. R. Arce, and G. Di Crescenzo, “Halftone
visual cryptography via error diffusion,” IEEE Trans.
Inf. Forensics Security, vol. 4, no. 3, pp. 383–396, Sep.
2009.
[18] L. A. MacPherson, “Grey Level Visual Cryptography for
General Access Structures,”Master Thesis, University
ofWaterloo,Waterloo, ON, Canada, 2002.
[19] Z. Zhou, G. R. Arce, and G. Di Crescenzo, “Halftone
visual cryptography,” in Proc. 2003 Int. Conf. Image
Processing, 2003, vol. 1, pp. I-521–I-524.
[20] D. Jin, W. Q. Yan, and M. S. Kankanhalli, “Progressive
color visual cryptography,” J. Electron. Imag., vol. 14,
no. 3, p. 033019, 2005.
[21] C. C. Lin and W. H. Tsai, “Visual cryptography for graylevel images by dithering techniques,” Pattern Recognit.
Lett., vol. 24, no. 1-3, pp. 349–358, 2003.
[22] J. O. Limb, “Design of dither waveforms for quantized
visual signals,” Bell Syst. Technol. J., vol. 48, no. 7, pp.
2555–2582, 1969.
[23] W. Zhou and A. C. Bovik, “A universal image quality
index,” IEEE Signal Process. Lett., vol. 9, no. 3, pp. 81–
84, Mar. 2002.
[24] Z. M. Wang and G. R. Arce, “Halftone visual
cryptography through error diffusion,” in IEEE Int.
Conf. Image Processing, 2006, pp. 109–112.
[25] Z.M.Wang, G. R. Arce, and G. Di Crescenzo, “Halftone
visual cryptography via direct binary search,” in Proc.
EUSIPCO’06, Florence, Italy, Sep. 2006.
http://www.ijettjournal.org
Page 3594