Data Security Utilizing a Memristive Coupled Neural Network in 3D Models

• Main Authors: Gabr, Mohamed, Diab, Amr, Elshoush, Huwaida T., Chen, Yen-Lin, Por, Lip Yee, Ku, Chin Soon, Alexan, Wassim
• Format: Article
• Published: Institute of Electrical and Electronics Engineers 2024
• Subjects: QA75 Electronic computers. Computer science
• Online Access: http://eprints.um.edu.my/47082/; https://doi.org/10.1109/ACCESS.2024.3447075

This article proposes a novel double data security algorithm that first encrypts sensitive data using a two-stage encryption method based on numerical solutions from a fractional-order memristive coupled neural network system. Solutions are obtained to generate encryption keys and construct S-boxes, which are then applied along with an initial key to encrypt the data bits through repeated XOR and S-box operations. The encrypted output is then hidden imperceptibly within 3D geometries by slightly modifying model points based on the encrypted data bits. This two-pronged approach provides enhanced protection for confidential information compared to single encryption or data hiding alone. Numerical experiments demonstrate the effectiveness of encryption in obscuring patterns while data extraction from modified 3D models validates recovery with negligible visual impact. Additionally, the proposed encryption scheme is shown to be superior to the standard AES-256 algorithm in terms of both computational efficiency and security against brute-force attacks. Through a synergistic blend of robust encryption and stealthy data hiding within 3D objects, the presented algorithm can reliably ensure privacy for sensitive digital data transmissions and storage applications.