Image Steganography

Dr Ian Cornelius

Hello

Hello (1)

Learning Outcomes

Understand the concept of image steganography
Demonstrate knowledge on how to use image steganography in a body of work

Steganography

Steganography (1)

Steganography is concerned with the study and practice of concealing information in objects
It is done in a manner that it deceives the viewer
- the viewer is under the impression there is no hidden content in the object
Essentially, the information is hidden in plain sight
- only the intended recipient will be able to view it

The term steganography is concerned with the practice of concealing information within an object. It has been around since 1499, where the first use of the term was by Johannes Trithemius in his steganographia, a treatise on cryptography and steganography, which was disguised as a book about magic.

The concealing of information is often done in such a way that it can deceive the viewer from believing that there is content hidden inside an object. The information is hidden in plain sight of the person viewing the object; however, only the intended recipient is able to view or extract that information.

Steganography (2)

Is this Cryptography?

No.
Cryptography is concerned with modifying a string to make it difficult to get the original string
- you are unable to get the original string back from the modified version
Both the original and modified string look completely different to one another
- i.e. abcd -> 1@z*

Image Steganography

Image Steganography (1)

A technique concerned with hiding data inside an image
Done in a manner that prevents an unintended user from detecting the hidden message
The following elements are required:
- a cover image: an image that will hold the message
- the message: the message to be sent, it can be plain or encrypted text or even an image
- a key: the key is used to hide the message, it is optional

Image steganography on the other hand is used to deceptively hide a message inside another message, and as such, modifying it. The modified message, however, is expected to look very similar to the other message. To perform a task such as this, we need the following:

a cover image which will hold the message you want to share with the other person
a message that you wish to send. The message can be plain-text, or it can be taken a stage further and encrypted. It could also be another image
a key, which is used to hide the message inside the image; this is an optional step and is not always required

Let’s have a look at an example. On the screen you can see an image of a cat, and it appears to have been duplicated. On the left-hand-side of the screen is the image of cat as it is; no hidden meaning or text. But with some magic and turning of the cogs, we can encode this image with a string, and the image on the right-hand-side of the screen is the same image as on the left. However, this image actually holds our hidden string.

But by looking at the image, there is no noticeable difference, and that is the whole purpose of image steganography. We want to be able to hide a message in plain-sight of other people, without them realising there is a message inside it. Now you do not know particularly what has been hidden inside this image. But hopefully, once you have completed your assignment, you can decode this image and I will supply you the encoded image upon request.

Image Steganography (2)

Applications of Image Steganography

Image steganography is useful for a multitude of things:
- securing private files
- transmitting messages or data with revealing the existence of a message
- hiding passwords or encryption keys
- transporting sensitive documents between users

Image steganography can be useful for a variety of different things. For example, hackers may use steganography techniques for malware transmission; alternatively, intelligence agencies can use them for communication between agents and staff.

In today’s world, attackers use Powershell or bash scripts to automate their attacks. For example, attackers can embed an actual script within a macro-enabled Excel spreadsheet or Word document, and once a victim opens this file, they will activate the embedded script. Potentially causing havoc on their machine. The attacker is using a steganography application to take advantage of common Windows applications and features such as Excel and PowerShell to perform their attack. All the victim is required to do is read the document and a series of events will unfold. This may involve the following steps:

First, the victim clicks on an Excel document that an attacker has modified using steganography.
Once clicked, it unleashes a hidden PowerShell script.
This script then installs an installer application into the Windows computer, and this installer moves quickly and is so subtle that a typical antivirus application does not notice it.
This downloader then goes out to the internet and grabs updated versions of malware such as URLZone (or more recent tools) that can then compromise the victim’s computer.

Image Steganography (3)

Types of Steganography Techniques

Several types and forms of steganography:
- physical: does not require the use of digital mediums or files; this type includes:
  - passing messages written with invisible ink which is read by the recipient by applying certain chemicals
  - use of ciphering techniques to hide the information with textual information, i.e. caesar cipher
- microdots: shrinking messages to tiny dimensions, becoming almost invisible
- digital: involves the use of digital mediums such as image, audio and video files

There are a variety of different types and forms of how steganography can be done. For this module, we are mostly concerned with using digital mediums such as images, audio and video files. However, physical mediums can be used. For example, you could write a message with invisible ink that can only be read by the recipient when using a certain type of chemical.

You may have done this when you were younger by using a wax candle to write a message and then using crayon to rub over the message to display the contents of what was written. However, this is not the only method; an alternative method is using a cipher technique such as a caesar cipher. You should already be familiar with this concept, but briefly, it is where each letter in the alphabet is shifting by a number to replace it with an alternative character. The other type of steganography is microdots, and this is concerned with shrinking messages to a tiny dimension whereby it comes almost invisible.

To hide information in each one of these mediums, a different technique is required. Each of these techniques will have their own advantages and disadvantages.

Image Steganography in Practice

Image Steganography in Practice (1)

Requires knowing about pixels and colour models
- re-visit the previous video if you need to revise
Say we have a pixel with values (0, 0, 255)
- red = 0
- green = 0
- blue = 255
- therefore, our pixel is blue
For an 8-bit system, a pixel can accommodate eight digits
- represented in a binary format
- the largest number in eight bits is: 11111111
  - this is equal to 255
- the smallest number in eight bits is: 00000000
  - this is equal to 0
Our RGB values in binary are:
- binaryRGB = (00000000, 00000000, 11111111)

We shall now move on and look at how we can go about using image steganography in practice. The first piece of coursework for this module will look at image steganography, and encoding a string into an image. Before we can begin, you need to know how an image is constructed, i.e. using pixels and colour models. If you have not already done so, I would recommend you watch the previous lecture video to this one. For the purpose of this lecture, when I refer to an image and its color model, we shall be using the red, green and blue (RGB) model.

Let’s say we have an image that consists of a single pixel. The RGB values for this pixel are: zero, zero, and two hundred and fifty-five. In this instance, the pixel will be blue as we have no intensity set in red or green. With a maximum intensity set in the blue bit. For an 8-bit system, we can convert our values into a binary format, which will consist of zeroes and ones. For example, two hundred and fifty-five in binary as an eight-bit is represented as: 1 1 1 1 1 1 1 1 and 0 0 0 0 0 0 0 0 for black. On the screen, you can see that I have converted our RGB value to the binary format.

Image Steganography in Practice (2)

Consider the following table, representing 3 pixels
Each pixel is a particular RGB value associated to it in eight-bit form

Pixel	RGB	R	G	B
1	(45, 28, 220)	00101101	00011100	11011100
2	(166, 196, 12)	10100110	11000100	00001100
3	(31, 86, 94)	00011111	01010110	01011110

When it comes to representing the colours of our pixels, each bit in the red, green and blue colour space needs to be converted to its binary representation. Let’s consider an image with only three pixels, each row in the table on the screen represents a pixel, with each bit from our RGB tuple represented as an eight-bit binary value.

Let’s have a look at the first row. The pixels RGB value is 45, 28 and 220, and we can see that this is a shade of purple. When we convert the decimal value for red to binary, we get the value: 0 0 1 0 1 1 0 1 and we can do this for each of the green and blue bits.

This process is then repeated for each other pixel in our image, in this case pixels two and three. The colour for each of these pixels is yellow and turquoise, respectively.

Image Steganography in Practice (3)

How can we hide the number 169 into it?
First, we need to convert the decimal number to binary: 10101001
Each digit of the binary number is then used to replace the least significant bit (LSB) from our pixels
- shown in the bold and red below

Pixel	RGB in Decimal	R	G	B
1	(45, 28, 221)	00101101	00011100	11011101
2	(166, 197, 12)	10100110	11000101	00001100
3	(30, 87, 94)	00011110	01010111	01011110

Some of our RGB values in decimal have also changed, due to changing the least significant bit
- shown in bold and blue

So how would we go about hiding a number in our image? Well, let’s consider we want to hide the number one hundred and sixty-nine. The first thing that we need to do is convert our decimal number into its binary equivalent, in this instance, our binary string is 1 0 1 0 1 0 0 1.

Each bit from our binary number is used to replace the least significant bit from our pixel. In this instance, it is the eighth-bit in our binary number for each of the RGB values in our pixels. To show this easier, in the table the bit that has been changed is bolded and changed to red.

When we change this least significant bit, our RGB decimal values may also be changed, and again this is represented in bold and blue. In this instance, we can see that our bolded, blue values have incremented by one for where we have changed a bit to 1. This is because the eighth-bit of a binary string is the value 1.

Image Steganography in Practice (4)

Before

Pixel	RGB in Decimal	R	G	B
1	(45, 28, 220)	00101101	00011100	11011100
2	(166, 196, 12)	10100110	11000100	00001100
3	(31, 86, 94)	00011111	01010110	01011110

After

Pixel	RGB in Decimal	R	G	B
1	(45, 28, 221)	00101101	00011100	11011101
2	(166, 197, 12)	10100110	11000101	00001100
3	(30, 87, 94)	00011110	01010111	01011110

Now let compare our pixels after we have changed the least significant bit. We can see straight away by looking at the image color (in the last column) that there is very little difference between the two images. That is because we have only changed the value in our RGB color space by an increment of one, and this does not make a great difference when it comes to color representation.

Therefore, in this instance, we have been able to encode the binary for our number 169, nice, into our image without causing much difference to the image/pixel itself. If we were to have some method of decoding this encoded image, we could easily extract our number once again. However, I will not talk you through this method, as there must be some element of work for your coursework.

Image Steganography in Practice (5)

The process we have just gone through is known as the Least Significant Bit (LSB)
- a common method that is often used for image steganography
Takes into account the pixel information of an image
Works best when the image file is larger than that of the message

LSB Algorithm Steps

Step 1. Select a cover image and choose a message to hide
Step 2. Find the pixels of the cover image
- 2a. Extract the RGB values of the first pixel
- 2b. Convert each value to its binary equivalent
Step 3. Extract the first character of the message
- 3a. Convert the character to its binary value
Step 4. Hide each digit of the characters binary value into the last bit of the RGB binary value
- 4a. Move onto the next pixel if required
Repeat Step 3 to 4a as necessary until all characters of the message are completed.

This whole process is known as least significant bit steganography, and this is the type of process I would expect students to follow for their coursework; albeit we will not be concerned with just numbers, but also letters and symbols. For example, I may want you to encode the module title into an image. However, to encode such a large string, we would also need an image with enough pixels to manipulate. This three-digit number took in total, just three pixels.

LSB is one of the methods used for image steganography and there are others. However, this is the simplest method. To help you understand this process better, and to assist you in your coursework, I have provided some steps in which should help you with the method of encoding a string into an image.

However, keep in mind that the coursework will also expect you to decode an image. In this instance, you may want to do some research on how you can go about reversing the encoding process. This is not taught in this lecture, as I expect students to do some self-guided study.

Goodbye

Goodbye (1)

Questions and Support

Questions? Post them on the Community Page on Aula
Additional Support? Visit the Module Support Page
Contact Details:
- Dr Ian Cornelius, ab6459@coventry.ac.uk