A fuzzy zeroing neural network and its application on dynamic Hill cipher

Cryptography is the core of information security, and the Hill cipher is one of the most important methods for cryptography. For the purpose of the improvement in the security of traditional Hill cipher (THC) with time-invariant key, a new dynamic Hill cipher (DHC) method with time-varying key is pr...

Celý popis

Uloženo v:

Podrobná bibliografie
Vydáno v:	Neural computing & applications Ročník 37; číslo 17; s. 10605 - 10619
Hlavní autoři:	Jin, Jie, Lei, Xiaoyang, Chen, Chaoyang, Lu, Ming, Wu, Lianghong, Li, Zhijing
Médium:	Journal Article
Jazyk:	angličtina
Vydáno:	London Springer London 01.06.2025 Springer Nature B.V
Témata:	Artificial Intelligence Color imagery Computational Biology/Bioinformatics Computational Science and Engineering Computer Science Cryptography Data Mining and Knowledge Discovery Encryption Image Processing and Computer Vision Invariants Neural networks Probability and Statistics in Computer Science Robustness (mathematics) S.I.: Timely Advances of Deep Learning with applications and Data Driven Modeling Security Special Issue on Timely Advances of Deep Learning with applications and Data Driven Modeling Hill cipher Encryption and decryption Robustness Zeroing neural network (ZNN) Fuzzy theory
ISSN:	0941-0643, 1433-3058
On-line přístup:	Získat plný text
Tagy:	Přidat tag Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!

Popis
Shrnutí:	Cryptography is the core of information security, and the Hill cipher is one of the most important methods for cryptography. For the purpose of the improvement in the security of traditional Hill cipher (THC) with time-invariant key, a new dynamic Hill cipher (DHC) method with time-varying key is proposed in this paper, which replaced the time-invariant matrix key of the THC with a time-varying matrix key for encryption. In order to ensure the efficiency and accuracy of the proposed DHC, the zeroing neural network (ZNN) is adopted to solve the time-varying key inversion (TVKI) matrix of the DHC decryption. Generally, conventional zeroing neural network (C-ZNN) models cannot effectively deal with the noises, and their convergence and robustness cannot be guaranteed in noisy environments. For the purpose of ensuring the robustness of the ZNN model for solving the TVKI matrix of the DHC decryption, a fuzzy activation function is designed, and a new fuzzy ZNN (FZNN) is constructed to solve the TVKI matrix of the DHC decryption. Comparative simulation results of the FZNN model with other C-ZNN models for solving the TVKI matrix of the DHC demonstrate that the FZNN model converges to the theoretical solution of the TVKI matrix within 0.8s in noisy environment, while other C-ZNN models take more than 0.8s or fail due to noises, which further validates the superior performances of the FZNN model. In addition, the encryption and decryption experiments on strings and RGB images are provided to prove the security of the proposed DHC, and comparative simulation results show that DHC has high feasibility and security in information encryption.
Bibliografie:	ObjectType-Article-1 SourceType-Scholarly Journals-1 ObjectType-Feature-2 content type line 14
ISSN:	0941-0643 1433-3058
DOI:	10.1007/s00521-024-10599-z