Reconfigurable security solution based on hopfield neural network for e-healthcare applications

Area of Science:

• Computer Science
• Biomedical Engineering
• Cryptography

Background:

• Securing medical images is critical for patient privacy in e-healthcare applications.
• Existing software-based encryption methods may not meet the performance demands of real-time medical imaging.

Purpose of the Study:

• To propose a novel, hardware-implemented encryption scheme for medical images.
• To leverage Field-Programmable Gate Arrays (FPGAs) for enhanced security and performance in e-healthcare.

Main Methods:

• A dual-layer encryption technique combining a Hopfield Neural Network (HNN) for diffusion and a stream cipher for confusion.
• Implementation on reconfigurable hardware (FPGA) for parallel processing and speed.
• Utilizing an image-specific key for enhanced security.

Main Results:

• The encryption scheme achieved high resilience against statistical attacks, with an average entropy of 7.99 and near-zero correlation.
• FPGA implementation demonstrated significant advantages in processing speed and parallelism.
• The FPGA implementation consumed 20% of hardware resources and 424.71 mW of power on an Intel Cyclone V FPGA.

Conclusions:

• The proposed FPGA-based encryption scheme offers a secure and efficient solution for medical image protection in e-healthcare.
• Hardware acceleration via FPGAs is beneficial for real-time, resource-constrained medical imaging applications.
• The dual-layer approach effectively enhances data confidentiality and integrity.