A study on color image encryption based on a new nonlinear dynamical confusion-diffusion algorithm model

Abstract To construct a chaotic image encryption algorithm resistant to plaintext attacks, a new two-dimensional chaotic map is proposed based on the characteristics of the Cubic and cosine functions. Additionally, a novel NL-PF (Nonlinear Parameter Feedback) pixel diffusion and scrambling method with a nonlinear dynamic model and parameter feedback is put forward, along with a mathematically rigorous proof of the algorithm. The encryption scheme includes NL-PF and DNA coding diffusion modules: first, NL-PF scrambling is used to diffuse pixel positions, and then encrypted images are generated using chaotic sequences based on DNA base coding. To verify the effectiveness in resisting plaintext attacks, this paper focuses on validating the performance of the proposed scheme through both statistical analysis and cryptographic analysis. Statistical analysis shows that the scheme has excellent resistance against statistical attacks; cryptographic analysis confirms that although this scheme does not possess formally cryptographically significant IND-CPA (Indistinguishability under Chosen Plaintext Attack) security, it exhibits good robustness in terms of statistics and diffusion when resisting Known Plaintext Attacks (KPA) and Chosen Plaintext Attacks (CPA).