Improving data protection in BSS based secure communication: mixing matrix design

Aslani, M. R ; Sharif University of Technology | 2021

154 Viewed
  1. Type of Document: Article
  2. DOI: 10.1007/s11276-021-02609-y
  3. Publisher: Springer , 2021
  4. Abstract:
  5. Abstract: In this paper, a secure and efficient Blind Source Separation (BSS) based cryptosystem is presented. The use of BSS in audio and image cryptography in wireless networks has attracted more attention. A BSS based cryptosystem consists of three main parts: secret data, secret keys, and mixing matrix. In this paper, we propose a new design to create a proper mixing matrix in BSS based cryptosystem. We offer a mathematical criterion to select mixing matrix elements before encryption. The proposed criterion gives a simple way to attach the secret sources to keys, which makes source separation very hard for the adversary. Versus, we show that using the random mixing matrix can lead to data security loss. The attacks used for security tests in this paper are "Differential Attack" and "Denoising Attack," which are among the most effective in this field. These attacks will apply to cryptosystems based on the random and the proposed mixing matrix. The visual results of the attacks in the experiments will show that the "proposed mixing matrix based cryptosystem" will be more secure than the "random mixing matrix based cryptosystem." We also used the correlation coefficient criterion to compare the two cryptosystems more accurately. According to the experiments of this paper, the "proposed mixing matrix based cryptosystem" vs. the "random mixing matrix based cryptosystem" was able to reduce the adversary's source extraction quality rate from about 76% to 16%, on average. Graphical Abstract: [Figure not available: see fulltext.] © 2021, The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature
  6. Keywords:
  7. Cryptography ; Matrix algebra ; Mixing ; Privacy by design ; Correlation coefficient ; Differential attacks ; Image cryptographies ; Mixing matrix ; Random mixing ; Secret data ; Security tests ; Source extraction ; Blind source separation
  8. Source: Wireless Networks ; Volume 27, Issue 7 , 2021 , Pages 4747-4758 ; 10220038 (ISSN)
  9. URL: https://link.springer.com/article/10.1007/s11276-021-02609-y