All of the image work we will encounter in the chapter will make use of the Python Image Library (PIL). To install the Python image libraries by using PIP on Linux, use the following command:

$ pip install PIL

If you are installing it on Windows, you may have to use the installers that is available at http://www.pythonware.com/products/pil/.

Just make sure that you get the right installer for your Python version.

It is worth noting that PIL has been superseded with a newer version PILLOW. But for our needs, PIL will be fine.

Images are created up by pixels, each of those pixels is made up of red, green, and blue (RGB) values (for color images anyway). These values range from 0 to 255, and the reason for this is that each value is 8 bits long. A pure black pixel would be represented by a tuple of (R(0), G(0), B(0)), and a pure white pixel would be represented by (R(255), G(255), B(255)). We will be focusing on the binary representation of the R value for the first recipe. We will be taking the 8-bit values and altering the right-most bit. The reason we can get away with doing this is that a change to this bit will equate to a change of less than 0.4 percent of the red value of pixel. This is way below what the human eye can detect.

Let's look at the script now, then we will go through how it works later on:

  #!/usr/bin/env python

from PIL import Image

def Hide_message(carrier, message, outfile):
    c_image = Image.open(carrier)
    hide = Image.open(message)
    hide = hide.resize(c_image.size)
    hide = hide.convert('1')
    out = Image.new('RGB', c_image.size)

    width, height = c_image.size

    new_array = []

    for h in range(height):
        for w in range(width):
            ip = c_image.getpixel((w,h))
            hp = hide.getpixel((w,h))
            if hp == 0: 
                newred = ip[0] & 254
            else: 
                newred = ip[0] | 1

            new_array.append((newred, ip[1], ip[2]))

    out.putdata(new_array)
    out.save(outfile)
    print "Steg image saved to " + outfile

Hide_message('carrier.png', 'message.png', 'outfile.png')

First, we import the Image module from PIL:

from PIL import Image

Then, we create our Hide_message function:

def Hide_message(carrier, message, outfile):

This function takes three parameters, which are as follows:

Next, we open the carrier and message images:

c_image = Image.open(carrier)
hide = Image.open(message)

We then manipulate the image that we are going to hide so that it's the same size (width and height) as our carrier image. We also convert the image that we are going to hide into pure black and white. This is done by setting the image's mode to 1:

hide = hide.resize(c_image.size)
hide = hide.convert('1')

Next, we create a new image and we set the image mode to be RGB and the size to be that of the carrier image. We create two variables to hold the values of the carrier images width and height and we setup an array; this array will hold our new pixel values that we will eventually save into the new image, as shown here:

out = Image.new('RGB', c_image.size)

width, height = c_image.size

new_array = []

Next comes the main part of our function. We need to get the value of the pixel we want to hide. If it's a black pixel, then we will set the LSB of the carriers red pixel to 0, if it's white then we need to set it to 1. We can easily do this by using bitwise operations that uses a mask. If we want to set the LSB to 0 we can AND the value with 254, or if we want to set the value to 1 we can OR the value with 1.

We loop through all the pixels in the image, and once we have our newred values, we append these along with the original green and blue values into our new_array:

    for h in range(height):
        for w in range(width):
            ip = c_image.getpixel((w,h))
            hp = hide.getpixel((w,h))
            if hp == 0: 
                newred = ip[0] & 254
            else: 
                newred = ip[0] | 1

            new_array.append((newred, ip[1], ip[2]))

    out.putdata(new_array)
    out.save(outfile)
    print "Steg image saved to " + outfile

At the end of the function, we use the putdata method to add our array of new pixel values into the new image and then save the file using the filename specified by outfile.

It should be noted that you must save the image as a PNG file. This is an important step as PNG is a lossless algorithm. If you were to save the image as a JPEG for instance, the LSB values won't be maintained as the compression algorithm that JPEG uses will change the values we specified.

We have used the Red values LSB for hiding our image in this recipe; however, you could have used any of the RGB values, or even all three. Some methods of steganography will split 8 bits across multiple pixels so that each bit will be split across RGBRGBRG, and so on. Naturally, if you want to use this method, your carrier image will need to be considerably larger than the message you want to hide.

So, we now have a way of hiding our image. In the following recipe, we will look at extracting that message.