Image Encryption Algorithms Based On Chaotic System And Discrete Cosine Stockwell Transform

As an important information carrier,digital images are widely utilized in people’s daily life.Therefore,it is particularly vital to ensure the secure and stable transmission of images in harsh network environments.In this dissertation,two secure and efficient image encryption algorithms based on chaotic system and discrete cosine Stockwell transform were designed.The main research work is as follows.A new image encryption scheme based on 6D non-degenerate discrete hyper-chaotic system,2D discrete cosine Stockwell transform and DNA-level modulus diffusion was presented.The robustness of the encryption algorithm against the chosen-plaintext attack was enhanced by generating the initial conditions of the chaotic systems with the SHA-512 hash function value of the plaintext image and the external key.The transmission burden was reduced by compressing the original image with the 2D discrete cosine Stockwell transform.Then random DNA encoding was performed on the compressed image to obtain the DNA image.To speed up the encryption,a DNA-level modulus diffusion algorithm was designed,which can scramble and diffuse the pixels at the same time.The ciphertext image was obtained by re-encrypting the diffused DNA image with the bit-level permutation and the improved global dynamic diffusion.The two high-dimensional chaotic systems introduced into the image encryption scheme effectively increase the key space,and make the image encryption algorithm resist the brute-force attack.It is shown that the proposed image encryption algorithm is feasible and secure.By combining the chaotic system and the improved gravity diffusion method,a double color images encryption algorithm was proposed.The algorithm has advantages such as large key space,sensitivity to plaintext images and secret key,and resistance to common attacks.A composite image of the two original images was compressed by the 2D discrete cosine Stockwell transform,and a Fibonacci transform was performed on the compressed image to scramble the positions of the pixels.Bit-level permutation and backward diffusion were performed in turn,and then the diffused image was processed into a 1D vector by Zigzag scanning.Aiming at the fragility of linear encryption methods,a non-linear random diffusion method was designed to enhance the security of the cryptosystem.The image after the previous processing step was confused by the index sequence of the random sequence generated by the chaotic system.To enhance the robustness of the encryption algorithm against the chosen-plaintext attack,an improved gravity diffusion was designed,which allows a specific encrypted pixel value to affect every pixel of the ciphertext image.It is demonstrated that the proposed double color images encryption scheme is more secure and robust.