Sie sind auf Seite 1von 3

ENCRYPTION AND DECRYPTION

USING CHAOTIC NEURAL NETWORKS


J Component Project
For
Soft Computing

SUBMITTED BY:

VIJYANT RAJ SINGH(17BIT0329)

PRATEEK RAJ(17BIT0301)
Abstract:
In recent years, more and more electronic services and devices are
available, like mobile phones and personal digital assistant PDAs,
also started to provide the additional functions like saving and
exchanging multimedia data. The prevalence of multimedia
technology has promoted digital images and videos to play a more
significant role than the traditional dull texts, which demands a
serious protection of users’ privacy and so the protection of
multimedia data is becoming very important. There is a lot of private
data which is stored in the form of text like credit card numbers,
passwords and a lot of other things. Even the photos we click are
made up of tiny pixels, each of which is a mixture of 3 numbers
which represent the RGB value of the colour of that pixel. So to
protect the privacy of people and prevent any unauthorized access
to their data, encryption of this data is very important. There are a
lot of ways in which the data can be encrypted. In this report, we are
going to discuss one of them – Using chaotic neural networks to
convert the data into something different so that it cannot be
converted to its original form without a predefined key.

Methodology:
The chaotic neural network can be used to encrypt digital signal. A
network is called a chaotic neural network if its weights and biases
are determined by a chaotic sequence. A chaotic sequence highly
depend upon the initial condition and parameters, x(0)=0.75 and µ
=2.6 are set. Difficult to encrypt and decrypt the data without
knowing the x (0) and µ. Depending upon the chaotic sequence a
weight matrix and bias matrix is obtained and the net input is
obtained.
Chaotic neural network offers greatly increase memory capacity.
Features of CNN:

 High Security
 No distortion
 Suitable for system integration
Algorithm:

1. Set the values of the parameters mew and x(0)


2. Convert the input to binary
3. Evolve the chaotic sequence x(1), x(2) … x(length(input)) as
x(i)=mew*x(i1)*(1-x(i-1)) and create b(0), b(1) …
b(length(input)) from x(1), x(2) … x(length(input)) by
generating a scheme so that b contains the binary
representation of x
4. Generate the weight matrix and bias matrix for each bit of
binary value of each value in b as:
1, 𝑖 = 𝑗 𝑎𝑛𝑑 𝑏
𝑊𝑗𝑖 = {−1, 𝑖 = 𝑗 𝑎𝑛𝑑 𝑏
0, 𝑖

0. , 𝑏

𝑡
5. Apply the activation function

D’=f(∑(wji*di) + thetai)
6. Convert to whole numbers from binary
7. The array obtained is the encrypted array
The same procedure is followed in decryption also.

Software:
We will be using “MATLAB SOFTWARE” for Execution of our Project.

Expected Result:
Using the randomness of CNN, we can encrypt the numerical data in such a way
that it becomes very difficult, if not impossible for someone who doesn’t have
the initial parameters to crack the key and get the original data from the
encrypted data. Number of attacks on the electronic data are increasing day by
day. But on the other side cryptography technique is also improving. By using
cryptography the probability of stealing information becomes a lot less. So,
cryptography plays major role in information securing. But using neural network
in cryptography also plays an important role. So, our future work involves the
implementation of cryptography using neural networks.

Das könnte Ihnen auch gefallen