Encrypting Images Using the Hill Cipher Under the Control of a Pseudorandom Vector

Abstract

The present research proposes a novel ciphering scheme that enhances the classical Hill cipher by integrating a pseudo-randomly binary vectors to control the encryption procedure. The method begins by vectorizing the plain picture and dividing it into sub-blocks containing a group of pixels. To achieve strong confusion, an inversible matrix is dynamically generated using two well-known and widely used chaotically maps in the cryptographic domain. This is combined with an affine transformation and multiple pseudo-random translation vectors. A diffusion step follows, applying a matrix that is not necessarily invertible to strengthen resistance against differential attacks and to withstand any known attack. Findings based on simulations conducted on a large number of images of various sizes confirm high security and strong performance.