Medical images encryption using two improved Vigenère laps separated by genetic crossover

This paper presents a new method for encrypting medical images based on an improvement of the classical Vigenère technique. This new approach employs new large substitution tables dynamically generated from several pseudo-random vectors. To eliminate any correlation between neighboring pixels in the original image, confusion is established with a pseudo-random vector generated from chaotic maps. Followed by two enhanced Vigenère rounds, using a strong diffusion function to protect the system against any differential attacks, interleaved by a genetic crossover deeply adapted to image encryption using a pseudo-random crossover table. Finally, a global permutation is applied to the output vector. Simulations are run on many images of different sizes and formats chosen arbitrarily to ensure that our approach is not exposed to known attacks.