A Novel Image Encryption Technique Using EEG S-Box and Dynamic Permutation for IoT Applications

This paper introduces a new image encryption method using the S-Box obtained from EEG datasets utilizing correlation and diffusion. A modified algorithm has been proposed for extracting binary information from EEG datasets, which is subsequently utilized for generating 64-bit keys. The generated keys and the proposed S-Box, along with the Lehmer random number generator-based dynamic permutation, are used for the encryption of (Formula presented.) grayscale test images. The generated 64-bit keys are tested using NIST SP 800-22 analysis for their randomness. The strength of the S-Box is estimated using parameters like bijectivity, nonlinearity, and strict avalanche criterion. The strength of the proposed encryption algorithm is evaluated using a chi-square test, peak signal-to-noise ratio, number of pixels change rate (NPCR), mean squared error, unified average changing intensity (UACI), and so on. The proposed design has been successfully implemented on a Virtex-7 FPGA, and an ASIC implementation has also been completed using 45-nm technology. When compared to the LUT–CLA–QTL architecture, it has been observed that there is almost a 40.21% reduction in the area occupied by the design on a silicon chip using 45-nm technology. This work gives a secure encryption of the data along with reduced hardware resources applicable for the resource constraint devices in IoT applications.