Referring to Wang and Li's 2019 paper titled, "A color image encryption algorithm based on Hopfield chaotic neural network".
They have used a 3D Hopfield Chaotic Neural Network to generate a N*N*3 matrix by taking 3 intial values. But when i try to implement it, the process is pretty time consuming compared to using chaotic maps like arnold, henon, etc... If the process is time expensive then why do we use it? Or there could be issues with the implementation?
Attached formulae used in HCNN.
tanh(x) could be piecewise replaced by three straight lines, y(x) = -1, y(x) = x, and y(x) = 1.
Regards,
Joachim
Priyansi Parida ,
If you have access the source code or the interpreter you can run the code and use a profiler to see which sections of the code are consuming too much time and debug if it is an implementation problem.
"If the process is time expensive then why do we use it? Or there could be issues with the implementation?"
Usually there can be trade offs between strength of encryption and its efficiency, but if the encryption is easily defeated it is not worth its efficiency.
here is my article which can be an answer of your question
Article Securing physical layer using new chaotic parametric maps
- Badges
- Science topic