Image encryption using modified perfect shuffle-based bit level permutation and learning with errors based diffusion for IoT devices

• Published in: Computers & electrical engineering 2022-05, Vol.100, p.107954, Article 107954
• Main Authors: Devipriya, M, Brindha, M
• Format: Article
• Language: English
• Subjects: Algorithms; Bit level permutation; Block permutation; Chaos theory; Diffusion barriers; Encryption; Learning with errors; Permutations; Piecewise linear map; Pixels; Tent map
• ISSN: 0045-7906; 1879-0755
• DOI: 10.1016/j.compeleceng.2022.107954

Image encryption is important for IoT devices which handle military, medical and other confidential images. In this paper, stream cipher based image diffusion is performed using tent chaotic map and block based encryption is done by taking four bytes every time to multiply the matrix formed using the secret key and error matrix. Pixel permutation is performed in two stages: block level pixel permutation and whole image permutation to resist the statistical attacks. Bit level permutation is performed with piecewise linear chaotic map using modified perfect shuffle technique to overcome differential attacks. Various statistical and differential analyses are performed on the encrypted image and the results obtained using these analyses convey its resistance to statistical and differential attacks. [Display omitted] •Plain image based key generated with MD5 hash algorithm to resist differential attack•Block based scrambling is deployed to improve the randomness in pixel distribution•Bit level permutation is employed to decrease the correlation between the pixels•Learning with errors based diffusion is added to increase the complexity of diffusion