The Application of the Linear Congruential Generator, the Single Neuron Model and Intel's MKL to Image Encryption

With the vast improvements in data transfer, the demand for effective image encryption techniques is constantly increasing. This paper presents a three-layer image cryptosystem. The first layer employs a Linear Congruential Generator S-box, the second layer is the Single Neuron Model, and the third layer is an Intel MKL S-box. The performance evaluation of the proposed cryptosystem reveals high resistance to many types of attacks at a low computing cost. Moreover, the proposed image cryptosystem is suitable for real-time applications. A comparison of the suggested cryptosystem to other cryptosystems shows that it is also superior to its counterparts from the literature.