A two layer chaotic network based image encryption technique

In recent years various chaos based cryptographic algorithms have been suggested to develop secure image encryption techniques but still many of them have low computing security as they have only good diffusion property but no confusion/substitution property making them more vulnerable to attacks. In this regard, this paper proposes a new secured approach for image encryption and decryption using two layer chaotic neural network (CNN). The two layers of network realize diffusion and substitution to make the cipher more secure. The logistic chaotic map is used to design the weights and the biases for the layers of network whose initial condition is derived under the control of an external key. Theoretical analysis and experimental results show that the proposed algorithm has good computing security for encryption.