A reconstruction scheme for secure self-verification of medical images incorporating spatiotemporal chaos encryption

Medical image transmission in telemedicine environments poses significant challenges to data integrity and confidentiality, and existing security solutions generally fail to balance high-strength encryption with high-precision self-recovery. To address this integration challenge, this paper proposes a medical image security self-verification reconstruction scheme that leverages a spatiotemporal chaotic system. This approach establishes a novel dual-layer security framework. For data integrity, it combines binary block embedding with permutation ordered binary encoding to achieve precise, pixel-level localization and self-recovery of tampering within the regions of interest (ROIs) of medical images. For data confidentiality, a novel nonlinear coupled spatiotemporal chaotic system driven by deep feature fingerprints enables authorized access control for sensitive data while generating high-entropy ciphertexts, significantly enhancing the confusion and diffusion capabilities of the encryption algorithm. Experimental simulations validate the framework’s exceptional performance: the tamper detection rate reaches 93.12%, with restored images achieving an SSIM value of 0.9908. Metrics including information entropy, pixel change rate, and uniform average change intensity all approach theoretical optimal values, fully demonstrating the algorithm’s outstanding resistance to statistical and differential attacks. This research establishes a highly secure, integrated medical image transmission solution that substantially enhances the reliability of telemedicine data.