Active Image Forgery Detection Using Cellular Automata - Cellular Automata in Image Processing and Geometry - page 139

Ta b l e 7 . 8 Evaluation of secret key sensitivity and its dependency to the original image's
changing

Image Manipulation Sum of Eigenvalues

(Red Layer)

Mean of Eigenvalues
(Red Layer)

Sara Original Image 51 16

10 Pixels Changed 43 11

20 Pixels Changed 38 9

Forest Original Image 67 21

10 Pixels Changed 81 15

20 Pixels Changed 73 14

Ta b l e 7 . 9 Comparison with non-CA algorithms

Developers Algorithm Data or items used Encryption level

Van Droogenbroeck
et al. [19]

Triple DES and IDEA External Logo high-level

Pommer et al. [13] Selective Encryption

of Wavelet-Packet

External Logo high-level

Krikor et al. [9] DCT and Stream Cipher Pseudo-Random bit

sequence. (External Key)

high-level

Anoop et al. [1] Transform Domains

and Stream Ciphers

Stream RC4 key values high-level

models proposed in 1D Cellular Automata Internal properties of

the input image

low-level

this chapter

7.6 Conclusion and Future Work

A cellular automata approach presented here for active image forgery detection.
In this chapter we have presented one scenario based on LU decomposition and
other scenario based on the SV decomposition with combination of one dimensional
cellular automata. The cellular automata rule generates a cipher key which can be
used to embed into the image. Both procedures need the original images to notice
forgery detection.
Seven different experimental validations have also been done. These experimen-
tal results obtained from the methods, specially the diffusion and true and false alert
clearly shown the performance and reliability of the models. In future work, we will
aim on two dimensional cellular automata forms into the suggested framework.