A novel plaintext-related dynamic DNA image encryption algorithm based on a 4D conservative chaotic system

Abstract Conservative chaotic systems can circumvent the flaws of dissipative systems, as they do not form chaotic attractors, thus enhancing their resilience to attacks. In this paper, a new non-Hamiltonian 4D conservative system is proposed, and the system’s conservativeness and chaotic properties are confirmed through the analysis of equilibrium points, chaotic flow projections, Poincaré sections, Lyapunov exponents, and Hamiltonian energy. Moreover, the system exhibits rich dynamical behaviors such as multistability, spectral entropy complexity, and offset-boosting. Building on this, a plaintext-related initial state generation algorithm using hash function is proposed, increasing sensitivity to plaintext. Also, a dynamic DNA algorithm is proposed by using plaintext information to control DNA operations. This ensures that the encryption operation is unique for each image. This algorithm enriches the encoding rules, thus enhancing the algorithm’s security. Experimental results show that the algorithm exhibits outstanding performance and strong resilience against various attacks.