Chapter 1

Introduction

In todays’ world, information security is of great concern. The data is secured by dif­
ferent encryption methods. After performing encryption, the plain data is converted into
cipher­text and then it is sent to the intended user. The intended user, then has to perform
a reverse procedure i.e. convert the cipher­text into plain text, the process is as shown in
figure 1.1 [1]. Even if the secret data is accessed by unauthorised user, that user should not
be able to interpret the meaning of that data. This is the basic aim of cryptography. There
are different techniques for encrypting data. But for encrypting images, there are very few
techniques, some of them are steganography and visual cryptography.

Figure 1.1: Overview of General Cryptography

In steganography, the secret image is hidden inside a cover image. Thus to embed the
secret image different techniques are available [2]. But with this, the appearance of cover
image changes. Such changes in image can easily attract the attention of eavesdropper and
eavesdropper will come to know that some secret message or image is embedded in the
cover image. Also, if the intended user receives such an image, user will have to use a

1
computer to analyse and extract the secret image from the cover image. Visual Cryptogra­
phy (VC) is a technique used for sharing a secret image. The secret image can be shared
with desired and specific people. The secret image is divided into different parts. These
parts are equal to the number of people among whom the secret image is to be shared. The
secret image is revealed only if all the desired number of people are present. Each person
among whom the secret image is shared, receives an encrypted version of the secret image
printed on transparent sheets. These transparent sheets are called as shares or transparen­
cies. The image visible on these transparencies is made up of random pixels of black and
white colour. From this transparency, no information about the secret image is revealed.
The transparency looks like as shown in figure 1.2.

Figure 1.2: Sample Share or Transparency

When the secret image is to be revealed, the desired number of transparencies should
be stacked on each other. When these transparencies are overlapped, the contrast of some
areas becomes low and for some areas remains same. This happens as per the objects in
the secret image. The difference in contrast is created among different areas of the stacked
transparencies, this helps the Human Visual System (HVS) to identify the secret image.
This is the basic principle of sharing secret image using traditional visual cryptography.
In this technique, if the number of people for revealing the secret are less than the desired
number of people, then the secret image is not revealed. Thus, any individual transparency
or transparencies less than the desired number do not reveal any information about the
secret image. A sample recovered secret image looks like as shown in figure 1.3.
The idea of sharing a secret was first conceptualized by Adi Shamir [3]. A technique
to divide a data in ‘n’ pieces such that the data can be reconstructed by the ‘n’ pieces, but
any ‘n ­ 1’ pieces of data cannot reconstruct the original data was developed. This tech­
nique was used in cryptographic systems for key management. Later on this algorithm was

RIRD, Satara, Shivaji University, Kolhapur 2

 Figure 1.3: Sample Recovered Secret Image

applied on binary images and from there on Visual Cryptography (VC) began to evolve.
Visual cryptography was first developed by Moni Naor and Adi Shamir [4]. In VC tech­
nique, the secret image is divided into number of pieces. These pieces are called as shares.
These shares are printed on transparent sheets, hence also called as transparencies. These
transparencies are then distributed among different participants. If there are ‘n’ participants
and if atleast ‘k’ participants are required to reveal the secret image, then such a technique
is called as (k, n) VC technique, where k < n. The biggest advantage of visual cryptog­
raphy is that, for decrypting the secret image computational power is not required. The
human visual system (HVS) is capable of decrypting the secret image from the stacked
transparencies. For a (k, n) Visual Cryptography Scheme (VCS), ‘n’ is the maximum num­
ber of participants in the scheme. ‘k’ is the minimum number of participants required for
revealing the secret. ‘m’ is the number of subpixels and ‘α’ is a relative difference in the
weight of black and white pixels, when the shares are overlapped. For (k, n) VCS, two
conditions must be followed for proper implementation.

1. Contrast Condition: The Hamming weight for the rows of the black and white ma­
trices should be same. When the shares more than ‘k’ are overlapped, the difference
in Hamming weight of black and white pixels is formed. This creates difference in
contrast and hence the secret image is revealed.

2. Security Condition: When shares less than ‘k’ are overlapped, then there should not
be difference in Hamming weights of black and white pixels.

For a traditional (k, n) VCS, there is a problem of pixel expansion i.e. a single pixel in
secret image is represented by more pixels in the shares. This affects the aspect ratio of the
recovered secret image. A group of pixels required for representing a single secret pixel

RIRD, Satara, Shivaji University, Kolhapur 3

1.1 Organization of the Thesis

is called subpixels. For instance, if for sharing a black secret pixel, the subpixels are [1
0 1 0]. Here, the pixel expansion is 4, because 4 pixels are required for sharing a single
secret pixel. Now, if these subpixels are arranged in a single row as shown in the figure
1.4(a), then the dimensions of recovered secret image will increase in horizontal direction
drastically. This will disturb the look and aspect ratio of the recovered secret image. To
avoid this problem, the subpixels are arranged in a square form rather than a straight line
as shown in figure 1.4(b). Hence, the dimensions of recovered secret image will increase
in both directions and the aspect ratio of the secret image will be maintained.

(a) Horizontal (b) Block

Figure 1.4: Subpixel arrangement formats

1.1 Organization of the Thesis

The thesis is organized into six chapters. Chapter 1. Introduction: It describes the concept
of visual cryptography and motivation to choose this area of research.
Chapter 2. Literature review: In this chapter, VC techniques are studied. In order to
find out the problems in VC, different techniques were studied. Starting with traditional
VC which deals with implementation of basic VC. To remove the problems present in tra­
ditional VC, techniques like extended VC, for general access structures where introduced.
These techniques improve security and contrast of recovered secret image. Techniques
to share multiple secrets and reduce pixel expansion are also studied. The possibility of
cheating in VC is also described. Some applications are discussed, in which VC can be
used effectively to improve the security of the system. Finally, various image quality met­
rics are studied to evaluate the quality of recovered secret images.
Chapter 3. Problem Definition: In this chapter the detailed problem statement and
objectives of research are described.
Chapter 4. System Description: In this chapter, the algorithms and general description

RIRD, Satara, Shivaji University, Kolhapur 4

1.1 Organization of the Thesis

about the techniques that are implemented are described. This chapter gives overview about
all the techniques that are implemented.
Chapter 5. System Design: This chapter describes the detailed design and implementa­
tion of the three techniques, first, for sharing two secrets, second, for improving contrast of
recovered secret image and third, to reduce the pixel expansion. This chapter also describes
the implementation of the authentication system.
Chapter 6. Results and Discussion: This chapter consists of the results obtained from
the techniques that are implemented in previous chapter. The recovered secret images are
evaluated using various image quality metrics. The meaning and significance of every
metric is also described in detail.
Chapter 7. Conclusion: This chapter includes the concluding remarks about the re­
search work. The research contributions are highlighted in this chapter. The further areas
of research in the area of VC are also described in this chapter.

RIRD, Satara, Shivaji University, Kolhapur 5