In this paper, a multi-image encryption scheme based on compressed sensing (CS) and deep learning in the optical gyrator domain is proposed. Firstly, multiple plaintext images are compressed by CS to obtain multiple measurements, and then the pixels of each measurement are scrambled by using a chaotic system. Secondly, the scrambled measurements are combined into a matrix and diffused by XOR operation with a chaotic matrix. Finally, the diffused matrix is encoded with a random phase and an optical gyrator transform to obtain a complex-valued matrix, and the amplitude of the complex-valued matrix is taken as the ciphertext. In decrypt, plaintext images are reconstructed from the CS measurements by a neural network, which achieves high reconstruction speed and quality compared with the traditional algorithm. Especially, the data amount of plaintext images can be compressed by up to 8 times while achieving high decryption quality. To our best knowledge, CS reconstruction algorithms based on deep learning is firstly used for image encryption. Moreover, the proposed scheme is highly robust against occlusion, noise, and chosenplaintext attack.

The security of medical image transmission in telemedicine is very important to patients’ privacy and health. A new asymmetric medical image encryption scheme is proposed. The medical image is encrypted by two spiral phase masks (SPM) and the lower–upper decomposition with partial pivoting, where the SPM is generated from the iris, chaotic random phase mask, and amplitude truncated spiral phase transformation. The proposed scheme has the following advantages: First, the iris is used for medical image encryption, which improves the security of the encryption scheme. Second, the combination of asymmetric optical encryption and three-dimensional Lorenz chaos improves the key space and solves the linear problem based on double-random phase encoding. Third, compared with other encryption schemes, the proposed scheme has advantages in occlusion attacks, key space, correlation, and information entropy. Numerical simulation and optical results verify the feasibility and robustness of the encryption scheme.

No abstract

The optical cryptosystem based on equal modulus decomposition (EMD) has attracted wide attention due to its remarkable anti-attack characteristics. In this paper, we propose a novel fully convolutional network model, which is an end-to-end deep learning method, to attack the EMD-based cryptosystem. The trained network model can retrieve plaintext after inputting many ciphertext-plaintext pairs and optimizing parameters. Numerical simulation results and analysis show that EMD-based cryptosystems by Fourier and Fresnel transforms are both vulnerable to our proposed method. Furthermore, the proposed network model can also successfully attack the interference-based cryptosystem. Compared with other methods, the proposed attack method has the advantages of shorter training time and stronger generalization ability. The proposed method provides a new approach for cryptoanalysis of cryptosystem based on EMD.