Mutated Cleavages of Images for Stealth Disclosure: A Hopfield Neural Network Attractor (HNNA) Approach

摘要

Multimedia data encryption is an ever-growing and challenging domain in the emerging digital world, with image encryption at the forefront to provide privacy for the user’s data. Image encryption aims to generate a stable noisy image irrespective of the input image. The current solution to this communication challenge is to use complex mathematics with a series of algorithms known as public-key cryptography. But the increasing power of computers and the ingenuity of hackers is open up more and more cracks in this mathematical armor. An ideal solution to this problem would provide provable information security, guaranteeing the safety of our personal information. In addition to it, the system must adapt itself to day to day miracles of the real world. Neural assisted techniques are the solutions to satisfy the above-said requirements. Hence, this paper proposes Hopfield neural attractor-assisted image encryption scheme in which combined tent maps are employed for the substitution process. Nowadays, a chosen-plaintext attack is a big challenge for encryption schemes. To resist the chosen plaintext attack, this scheme proposes the higher bitplane confusion using the tent map, which replaces the conventional diffusion process. Besides, the proposed scheme offers desire metrics such as the average correlation of 0.003, entropy as 7.99, and keyspace of 10184.