Dynamic thresholded chest medical image encryption scheme based on 2D-ESLMM chaos

Abstract With the increasing use of telemedicine and cross-hospital data exchange, chest medical images containing sensitive diagnostic information face growing risks of unauthorized access, tampering, and data leakage during transmission. Therefore, developing an efficient and secure encryption scheme is of great importance for protecting patients’ privacy and ensuring the reliability of medical data communication. This paper presents a chest image encryption method based on an improved 2D-ESLMM (2D-Enhanced Sine Logistic Modulation Map) chaotic system to address the limited parameter range, instability, and periodicity in traditional maps. By integrating exponential functions and dynamic parameters, the chaotic behavior and randomness are significantly enhanced, as verified by phase diagrams, bifurcation plots, Lyapunov exponents, and 0–1 tests. The proposed system acts as a key generator combined with a dynamic threshold confusion mechanism based on regions of interest (ROI) and a bidirectional feedback diffusion strategy. This configuration serves to enhance global obfuscation and improve resistance to differential attacks. Experimental results confirm that the scheme expands key space, increases entropy, reduces pixel correlation, and achieves NPCR and UACI values close to theoretical ideals, demonstrating strong security and robustness.