A Unique Technique to RGB Image Cryptography via Design of Deep Learning
Abstract
Images are either shared or saved in the current digital era. Protecting photos from unwanted access or alteration is a conscious need. The suggested work is an attractor-inspired deep learning for picture encryption, which is a non-linear dynamic system that makes it easier to establish a consensus independent of the beginning and boundary conditions of the system. To describe the chaotic behaviour, the Lorenz 3D chaotic Map solution is used as the attractor. This attractor transitions into a predictable set of solutions with regular behaviour, which makes decrypting the cipher picture easier and quicker. All of the researches on encryption are on achieving high resistance to differential attacking metrics in order to demonstrate the encryption's strength; however, our work focuses primarily on data at rest and optimises computations while protecting pictures from manipulation.