Abstract
A new image encryption algorithm is proposed by combining fractional Hermite transform (FHT) with hyper-chaos map. In the first round of pixel scrambling, the original image is performed by the Zigzag operation and the Arnold transform. Then the scrambled image matrix is XORed with an encryption sequence composed of an 8-bit linear feedback shift register. Compared with the traditional Hermite transform, the FHT has better noise tolerance, and it could process the non-stationary input signals better. Subsequently, the encrypted data is re-encrypted by the cyclic shift operation based on the hyper-chaotic system. The four initial values of the hyper-chaotic system are obtained by combining plaintext image with SHA-512. Finally, the processed image matrix is XORed with a chaotic matrix generated by the three-dimensional Logistic-sine-coupling map. The presented image encryption algorithm has a large key space and strong anti-noise ability. Simulation results show that the proposed algorithm is feasible, secure and effective.
Acknowledgments
This work was supported by the National Natural Science Foundation of China (grant number 61861029), the Major Academic Discipline and Technical Leader of Jiangxi Province (grant number 20162BCB22011), and the Natural Science Foundation of Jiangxi Province (grant number 20171BAB202002).
Disclosure statement
No potential conflict of interest was reported by the author(s).