Invertible encryption network for optical image cryptosystem

Abstract In this paper, an invertible encryption network (IENet) is firstly proposed for optical image cryptosystem, which applies a novel learning way to optical image encryption for the first time, to the best of our knowledge. Here, the quadratic phase and double phase method are firstly applied to generate the phase-only hologram, and then it is encrypted by IENet. The ciphertext is finally obtained by learning the probability distribution from the most uniform matrix with an entropy of 8.0. Notably, the decryption process is the inverse process of the encryption. Due to the invulnerability of the ciphertext-only attack, chosen-ciphertext attack, and cryptosystem leakage, a high security is achieved by the proposed encryption method. Moreover, the optical video decryption is accomplished by combining optics and deep learning owing to the fast decryption of the neural network. The effectiveness and superiority of the proposed encryption method is verified by the simulation and experiment results. The proposed method provides a new approach and perspective for the learning-based encryption research.