1.

Introduction
Since the rise of the Internet one of the most important factors of information technology and communication
has been the security of information. Cryptography was created as a technique for securing the secrecy of
communication and many different methods have been developed to encrypt and decrypt data in order to keep
the message secret. Unfortunately it is sometimes not enough to keep the contents of a message secret, it may
also be necessary to keep the existence of the message secret. The technique used to implement this, is called
steganography.
Steganography is the art and science of invisible communication. This is accomplished through hiding
information in other information, thus hiding the existence of the communicated information. The word
steganography is derived from the Greek words “stegos” meaning “cover” and “grafia” meaning “writing” [1]
defining it as “covered writing”. In image steganography the information is hidden exclusively in images.
The idea and practice of hiding information has a long history. In Histories the Greek historian Herodotus writes
of a nobleman, Histaeus, who needed to communicate with his son-in-law in Greece. He shaved the head of one
of his most trusted slaves and tattooed the message onto the slave’s scalp. When the slave’s hair grew back the
slave was dispatched with the hidden message [2]. In the Second World War the Microdot technique was
developed by the Germans. Information, especially photographs, was reduced in size until it was the size of a
typed period. Extremely difficult to detect, a normal cover message was sent over an insecure channel with one
of the periods on the paper containing hidden information [3]. Today steganography is mostly used on
computers with digital data being the carriers and networks being the high speed delivery channels.
Steganography differs from cryptography in the sense that where cryptography focuses on keeping the contents
of a message secret, steganography focuses on keeping the existence of a message secret [4]. Steganography and
cryptography are both ways to protect information from unwanted parties but neither technology alone is perfect
and can be compromised. Once the presence of hidden information is revealed or even suspected, the purpose of
steganography is partly defeated [4]. The strength of steganography can thus be amplified by combining it with
cryptography.
Two other technologies that are closely related to steganography are watermarking and fingerprinting [5]. These
technologies are mainly concerned with the protection of intellectual property, thus the algorithms have different
requirements than steganography. These requirements of a good steganographic algorithm will be discussed
below. In watermarking all of the instances of an object are “marked” in the same way. The kind of information
hidden in objects when using watermarking is usually a signature to signify origin or ownership for the purpose
of copyright protection [6]. With fingerprinting on the other hand, different, unique marks are embedded in
distinct copies of the carrier object that are supplied to different customers. This enables the intellectual property
owner to identify customers who break their licensing agreement by supplying the property to third parties [5].
In watermarking and fingerprinting the fact that information is hidden inside the files may be public knowledge
– sometimes it may even be visible – while in steganography the imperceptibility of the information is crucial
[4]. A successful attack on a steganographic system consists of an adversary observing that there is information
hidden inside a file, while a successful attack on a watermarking or fingerprinting system would not be to detect
the mark, but to remove it [5].
Research in steganography has mainly been driven by a lack of strength in cryptographic systems. Many
governments have created laws to either limit the strength of a cryptographic system or to prohibit it altogether
[7], forcing people to study other methods of secure information transfer. Businesses have also started to realise
the potential of steganography in communicating trade secrets or new product information. Avoiding
communication through well-known channels greatly reduces the risk of information being leaked in transit [8].
Hiding information in a photograph of the company picnic is less suspicious than communicating an encrypted
file.
This paper intends to offer a state of the art overview of the different algorithms used for image steganography
to
illustrate the security potential of steganography for business and personal use. After the overview it briefly
reflects on the suitability of various image steganography techniques for various applications. This reflection is
based on a set of criteria that we have identified for image steganography. The remainder of the paper is
structured as follows: Section 2 gives the reader an overview of steganography in general and differentiates
between different kinds of steganography. In section 3 the most popular algorithms for image steganography are
discussed and compared in section 4. In Section 5 a conclusion is reached.
2. Overview of Steganography
To provide an overview of steganography, terms and concepts should first be explained. An overview of the
different kinds of steganography is given at a later stage.
2.1 Steganography concepts
Although steganography is an ancient subject, the modern formulation of it is often given in terms of the
prisoner’s problem proposed by Simmons [9], where two inmates wish to communicate in secret to hatch an
escape plan. All of their communication passes through a warden who will throw them in solitary confinement
should she suspect any covert communication [10].
The warden, who is free to examine all communication exchanged between the inmates, can either be passive or
active. A passive warden simply examines the communication to try and determine if it potentially contains
secret information. If she suspects a communication to contain hidden information, a passive warden takes note
of the detected covert communication, reports this to some outside party and lets the message through without
blocking it. An active warden, on the other hand, will try to alter the communication with the suspected hidden
information deliberately, in order to remove the information [5].
2.2 Different kinds of steganography
Almost all digital file formats can be used for steganography, but the formats that are more suitable are those
with a high degree of redundancy. Redundancy can be defined as the bits of an object that provide accuracy far
greater than necessary for the object’s use and display [11]. The redundant bits of an object are those bits that
can be altered without the alteration being detected easily [5]. Image and audio files especially comply with this
requirement, while research has also uncovered other file formats that can be used for information hiding.
Figure 1 shows the four main categories of file formats that can be used for steganography.
Figure 1: Categories of steganography
Hiding information in text is historically the most important method of steganography. An obvious method was
to hide a secret message in every nth letter of every word of a text message. It is only since the beginning of the
Text Images Audio/
video
Protocol
Internet and all the different digital file formats that is has decreased in importance [1]. Text steganography
using digital files is not used very often since text files have a very small amount of redundant data.
Given the proliferation of digital images, especially on the Internet, and given the large amount of redundant bits
present in the digital representation of an image, images are the most popular cover objects for steganography.
This paper will focus on hiding information in images in the next sections.
To hide information in audio files similar techniques are used as for image files. One different technique unique
to audio steganography is masking, which exploits the properties of the human ear to hide information
unnoticeably. A faint, but audible, sound becomes inaudible in the presence of another louder audible sound [1].
This property creates a channel in which to hide information. Although nearly equal to images in steganographic
potential, the larger size of meaningful audio files makes them less popular to use than images [8].
The term protocol steganography refers to the technique of embedding information within messages and
network
control protocols used in network transmission [13]. In the layers of the OSI network model there exist covert
channels where steganography can be used [12]. An example of where information can be hidden is in the
header of a TCP/IP packet in some fields that are either optional or are never used. A paper by Ahsan and
Kundur provides more information on this [13].
3. Image steganography
As stated earlier, images are the most popular cover objects used for steganography. In the domain of digital
images many different image file formats exist, most of them for specific applications. For these different image
file formats, different steganographic algorithms exist.
3.1 Image definition
To a computer, an image is a collection of numbers that constitute different light intensities in different areas of
the image [14]. This numeric representation forms a grid and the individual points are referred to as pixels.
Most images on the Internet consists of a rectangular map of the image’s pixels (represented as bits) where each
pixel is located and its colour [15]. These pixels are displayed horizontally row by row.
The number of bits in a colour scheme, called the bit depth, refers to the number of bits used for each pixel [16].
The smallest bit depth in current colour schemes is 8, meaning that there are 8 bits used to describe the colour of
each pixel [16]. Monochrome and greyscale images use 8 bits for each pixel and are able to display 256
different colours or shades of grey. Digital colour images are typically stored in 24-bit files and use the RGB
colour model, also known as true colour [16]. All colour variations for the pixels of a 24-bit image are derived
from three primary colours: red, green and blue, and each primary colour is represented by 8 bits [14]. Thus in
one given pixel, there can be 256 different quantities of red, green and blue, adding up to more than 16-million
combinations, resulting in more than 16-million colours [16]. Not surprisingly the larger amount of colours that
can be displayed, the larger the file size [15].

Table shows a comparison of the results of the metrics of distortion [9] (Average Absolute Difference, Mean
Squared Error, Lp-Norm, Laplacian Mean Squared Error, Signal to Noise Ratio, Peak Signal to Noise Ratio,
Normalised Cross-Correlation and Correlation Quality) applied to steganographic images obtained by different
tools. Obviously, a lower distortion represents a better steganographic method because it is closer to the values
 Resolution SNR PSNR MSE Distortion

Original image 600X800 of the

0,0 0,0 0,000 Nil original
1024X786
LSB image.
algorithm(1
600X800 7,5 2,1 0,020 12%
bit/pixel)

LSB algorithm
600X800,
(2bits/pixel) 4,0 6,7 0,043 15%
1024X786
LSB
algorithm(3
1024X786 2,3 3,2 0,137 20%
bits/pixel)

3.2 Image Compression

When working with larger images of greater bit depth, the images tend to become too large to transmit over a
standard Internet connection. In order to display an image in a reasonable amount of time, techniques must be
incorporated to reduce the image’s file size. These techniques make use of mathematical formulas to analyse
and condense image data, resulting in smaller file sizes. This process is called compression [15].
In images there are two types of compression: lossy and lossless [1]. Both methods save storage space, but the
procedures that they implement differ. Lossy compression creates smaller files by discarding excess image data
from the original image. It removes details that are too small for the human eye to differentiate [15], resulting in
close approximations of the original image, although not an exact duplicate. An example of an image format
that uses this compression technique is JPEG (Joint Photographic Experts Group) [14].
Lossless compression, on the other hand, never removes any information from the original image, but instead
represents data in mathematical formulas [15]. The original image’s integrity is maintained and the
decompressed image output is bit-by-bit identical to the original image input [1]. The most popular image
formats that use lossless compression is GIF (Graphical Interchange Format) and 8-bit BMP (a Microsoft
Windows bitmap file) [14].
Compression plays a very important role in choosing which steganographic algorithm to use. Lossy
compression techniques result in smaller image file sizes, but it increases the possibility that the embedded
message may be partly lost due to the fact that excess image data will be removed [7]. Lossless compression
though, keeps the original digital image intact without the chance of lost, although is does not compress the
image to such a small file size [14]. Different steganographic algorithms have been developed for both of these
compression types and will be explained in the following sections.
3.3 Image and Transform Domain
Image steganography techniques can be divided into two groups: those in the Image Domain and those in the
Transform Domain [2]. Image – also known as spatial – domain techniques embed messages in the intensity of
the pixels directly, while for transform – also known as frequency – domain, images are first transformed and
then the message is embedded in the image [20].
Image domain techniques encompass bit-wise methods that apply bit insertion and noise manipulation and are
sometimes characterised as “simple systems” [17]. The image formats that are most suitable for image domain
steganography are lossless and the techniques are typically dependent on the image format [18].
Steganography in the transform domain involves the manipulation of algorithms and image transforms [17].
These methods hide messages in more significant areas of the cover image, making it more robust [4]. Many
transform domain methods are independent of the image format and the embedded message may survive
conversion between lossy and lossless compression [18].