DNA coding and chaotic mapping boost performance in image encryption
DNA coding and chaotic mapping boost performance in image encryption lead image
Digital images are easily subject to a wide range of cyberattacks, including noise attacks, shear attacks, and entropy attacks. As image decryption technology progresses, single low-dimensional chaotic systems and simple “scrambling-diffusion” encryption algorithms can no longer keep images safe.
Tang et al. developed a new image encryption algorithm that draws on genetic random coding and chaotic mapping. Drawing on previously established four-dimension chaotic systems, their algorithm uses the SHA-256 ZigZag transform as its initial value to create a hyperchaotic system. Simulations show that it improves safeguards by increasing the correlation between the key and the plaintext as well as the randomness of the encryption process.
“At present, there are relatively mature research results on encryption methods that combine chaotic systems and DNA,” said author Chengwei Tang. “However, there are disadvantages such as poor resistance to chosen plaintext attacks, small ciphertext entropy, and a single DNA coding rule.”
The ZigZag transform, a two-dimensional matrix based on the Chaotic Sequence of Chebyshev to obtain the scrambled and diffused ciphertext image, facilitates increasing the complexity of encryption by rearranging the image’s matrix and reducing the correlation between adjacent pixels.
The correlation coefficients of adjacent pixels of the encrypted images are all near 0, indicating the algorithm enhances the correlation between the key and the plaintext and the randomness of the encryption process, effectively improving the anti-attack ability. The information entropy is more significant than 7.997, relative to the theoretical value of 8.
The authors hope their work stimulates further work in maturing DNA and genetic algorithms.
Source: “Encryption algorithm based on improved four-dimensional chaotic system and dynamic DNA encoding,” by Chengwei Tang, Shibing Wang, Yubing Shu, and Fujun Ren, AIP Advances (2024). The article can be accessed at https://doi.org/10.1063/5.0207225 .
