big S-Boxes for Deeply Improved Hill Image Coding

Abstract

This paper introduces a novel approach to encrypt color images, utilizing multiple chaotic maps and working with any size pixel blocks. The encryption procedure commences by converting the original image into a vector, which is then divided into equal dimension subblocks. These blocks undergo a new Vigenère transformation, employing two large dynamic substitution tables interconnected by an enhanced Hill operation using a large invertible matrix generated from a pseudo-random table, controlled by another binary dynamics table. Additionally, the truncated block undergoes the same encryption process using another pseudo-random invertible matrix. To evaluate the effectiveness and resilience of our algorithm, we conducted numerous simulations on randomly selected images of varying sizes and formats from diverse databases.