Image Encryption Using Differential Evolution Approach in Frequency Domain

This paper presents a new effective method for image encryption which employs magnitude and phase manipulation using Differential Evolution (DE) approach. The novelty of this work lies in deploying the concept of keyed discrete Fourier transform (DFT) followed by DE operations for encryption purpose. To this end, a secret key is shared between both encryption and decryption sides. Firstly two dimensional (2D) keyed discrete Fourier transform is carried out on the original image to be encrypted. Secondly crossover is performed between two components of the encrypted image, which are selected based on Linear Feedback Shift Register (LFSR) index generator. Similarly, keyed mutation is performed on the real parts of a certain components selected based on LFSR index generator. The LFSR index generator initializes it seed with the shared secret key to ensure the security of the resulting indices. The process shuffles the positions of image pixels. A new image encryption scheme based on the DE approach is developed which is composed with a simple diffusion mechanism. The deciphering process is an invertible process using the same key. The resulting encrypted image is found to be fully distorted, resulting in increasing the robustness of the proposed work. The simulation results validate the proposed image encryption scheme.