Abstract
Many encryption systems face two problems: the key has nothing to do with the plaintext; only a single chaotic sequence is adopted during the encryption. To solve the problems, this paper proposes an image encryption method based on Hopfield neural network and bidirectional flipping. Firstly, the plaintext image was segmented into blocks, the resulting image matrix was block scrambled, and each block was bidirectionally flipped to complete the scrambling process. After that, the plaintext image was processed by the hash algorithm to obtain the initial values and control parameters of the chaotic system, producing a pseudo-random sequence. Then, a diffusion matrix was generated through the optimization by Hopfield neural network and used to derive a ciphertext image through diffusion transformation. Experimental results show that our algorithm is highly sensitive to plaintext, strongly resistant to common attacks, and very efficient in encryption.
Highlights
To solve the above defects, this paper explores key generation and image encryption strategy and proposes a chaotic image encryption algorithm based on Hopfield neural network and the image scrambling approach of bidirectional flipping
The initial values of logistic mapping were determined as x1 (0) 0.8761 and x2 (0) 0.7323; the control parameters of logistic mapping were finalized as r1 3.9695 and r2 3.8925; the total number of iterations (TNI) was set to 200; different hash arrays H were generated from different plaintext images
This paper mainly designs a chaotic image encryption algorithm based on Hopfield neural network and bidirectional flipping, a scrambling strategy
Summary
The safety of digital images, an important carrier of information, has attracted much interest and concern [1]. To ensure the safety of image information, it is highly necessary to develop a good encryption algorithm [2]. The chaotic system lays a good foundation for encryption systems, due to its excellent sensitivity to initial values. Some scholars presented encryption algorithms based on existing chaotic systems, focusing on the design of encryption strategies [9–16]. To solve the above defects, this paper explores key generation and image encryption strategy and proposes a chaotic image encryption algorithm based on Hopfield neural network and the image scrambling approach of bidirectional flipping. The plaintext image was segmented into multiple N × N blocks, and the resulting image matrix was block scrambled. Multiple sequences were taken as the initial conditions of Hopfield chaotic neural network, which creates the key flow of the diffusion matrix. The safety and reliability of our algorithm were demonstrated by comparing it with similar algorithms developed since 2017
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
