Medical image encryption algorithm based on a novel 4D chaotic system and dynamic zigzag transformation

Abstract In medical imaging, where diagnostic integrity and data confidentiality must be preserved simultaneously, traditional cryptographic approaches often fail to achieve both high security and lossless recovery. This paper proposes a lightweight and reversible medical image encryption algorithm based on a novel four-dimensional chaotic system to address the critical requirements of both security and diagnostic integrity in healthcare communications. The initial parameters of the chaotic system are derived from the SHA-256 hash of the plaintext image, ensuring plaintext-dependent key generation. The encryption process integrates a dual-stage permutation-diffusion framework: permutation combines global bit-level row-column shifting with intra-block dynamic Zigzag scanning, while diffusion employs chaotic sequences and inter-block feedback for comprehensive pixel transformation. Extensive experiments on various medical images demonstrate that the method achieves near-ideal statistical performance, with information entropy close to 8, NPCR around 99.6067%, and UACI approximately 33.4629%, alongside strong robustness against noise, cropping, and plaintext attacks while preserving lossless reversibility. Comparative analysis confirms that the proposed scheme achieves superior or comparable security to recent methods, making it a practical and efficient solution for securing medical images in telemedicine and IoMT applications.