An asymmetric cryptosystem based on generalized singular value decomposition with QZ modulation, using Fourier domain

In this paper, a cryptosystem with an asymmetric approach and using Fourier transform is presented. The presented scheme is based on two matrix algorithms namely generalized singular value decomposition (GSVD) and QZ algorithm. Random phase masks (RPM) used in the cryptosystem as encryptions keys while the keys generated by QZ modulation and GSVD are treated as private keys. The decryption of image is possible only using the correct decryptions keys. It is difficult to retrieve the plaintext image information if any one of the private keys is missing. Simulation results obtained using MATLAB(R2022a) confirm the validity of the cryptosystem by evaluating the key sensitivity, and resilience against occlusion-, and noise attack. Value of the mean-square-error (MSE) between the original and deciphered image is evaluated to confirm the system's efficacy.