Active Image Forgery Detection Using Cellular Automata - Cellular Automata in Image Processing and Geometry - page 127

rule could be applied to boundary cells. Since we are facing with the real input val-
ues (i.e. mean and range) for our proposed cellular automata, and we intend to apply
XOR logical operation as a local rule, so we have to convert them to binary values.
We foremost convert from real to integer values to avoid existing alteration com-
plexities and limitations, applying recursive integer to binary alteration algorithm.
Finally, we embed the output of the proposed cellular automata into the LSB
(Least Significant Bit) of the first eight pixels in the original image. LSB substitu-
tion is the process of modifying the least significant bit of the pixels of the input
image. By doing this process, the value of a pixel is changed slightly, so the changes
are not reflected physically in the output image. The proposed algorithm has been
applied on grayscale images with pixel values between 0 and 255 in which the spe-
cific statistical information could be in the same range. Therefore, a byte should be
enough to store output value as a cipher key. The experiments indicate that the LSB
usage of the first eight pixels in this step is almost proper to obtain a well result on
both time consuming and PSNR.
Here we briefly describe the statistical operations which are used in the proposed
cellular automata.
Fig. 7.2 Block diagram of the proposed method
Fig. 7.3 Block diagram of the proposed cellular automata for cipher key creation