Parallel AI-driven framework for post-quantum secure medical image communication using swin-transformer restoration.

Area of Science:

• Medical Imaging
• Telemedicine
• Network Communication

Background:

• Medical image transmission faces challenges like packet loss, noise, and privacy risks.
• Traditional error correction and encryption methods are insufficient for high-resolution medical imaging.

Purpose of the Study:

• To propose a next-generation AI-assisted framework for privacy-preserving medical image transmission.
• To integrate advanced AI, neural coding, and post-quantum cryptography for enhanced reliability and security.

Main Methods:

• Utilized a Restormer/Swin-Transformer hybrid network for image corruption detection and restoration.
• Incorporated Deep Joint Source-Channel Coding (DeepJSCC) and Neural Error Correction Codes (NECC) for transmission robustness.
• Implemented lattice-based post-quantum cryptography (CRYSTALS-Kyber, CRYSTALS-Dilithium) for enhanced security.

Main Results:

• The framework demonstrated superior performance in recovering corrupted medical image regions.
• Achieved enhanced transmission robustness and cryptographic security compared to traditional methods.
• Enabled real-time processing via GPU acceleration and federated learning for privacy-preserving collaborative training.

Conclusions:

• The proposed AI-assisted framework significantly improves medical image reconstruction fidelity, transmission robustness, and security.
• Offers a robust solution for privacy-preserving medical image communication in distributed healthcare systems.
• Addresses the limitations of traditional methods in modern high-resolution medical imaging environments.