Research On Optical Encryption And Authentication Technology Combined With Neural Network

In the information age,frequent information exchange has become a part of the daily life of the public,so information security has been highly concerned by people.Optical information security technology is widely used in information encryption,information hiding and authentication systems because of its high parallelism and the speed of light processing of large capacity optical information.At the same time,with the rapid development of hardware technology,major breakthroughs have been made in the research of artificial intelligence,machine learning,neural network and other fields.These research results have penetrated into other fields and solved many problems difficult to be solved by traditional technology.Convolutional neural networks have been widely used in optical information security due to their ability to extract image features,including optical information authentication,attack on encryption systems and design of optical encryption systems with higher security.However,the neural network running on the computer will still be limited by the hardware level.In order to break through this limitation,the scheme of all-optical neural network was proposed in 2018.It combines the advantages of optical system and neural network,which can provide a new tool for the field of optical information security and open a new chapter for information security technology.Based on the above research background,this paper carries out the application of neural network technology in the field of optical information security.In this paper,the mathematical principle and development of neural networks are introduced systematically,and the architecture and training process of fully connected neural networks and convolutional neural networks are introduced in detail.At the same time,the principle of information optics is briefly introduced and several typical optical encryption systems are given.Then,this paper proposes a hidden image encryption system based on neural network,which encodes the hidden image placed in phase information and adds it to the host image weighted by double random phase principle.Using conventional decryption method and attack method based on phase recovery algorithm to attack this system can not get the image information,is a conventional irreversible coding system.However,by using convolutional neural networks to deal with a large number of ciphertexts and their corresponding plaintexts,a neural network model can be trained,which can directly convert ciphertexts into corresponding plaintexts,and successfully decrypt conventional irreversible encryption systems.And the decryption process of the system does not need to provide the host image,it is a blind watermark decryption technology.In addition,this paper proposes an optical authentication system based onD~2 NN.After the beam passes through the public key and private key,an image to be authenticated is generated at a specific distance.The image can be entered into the authentication system for security authentication.Because the system uses invisible terahertz light,it has the characteristics of concealment and security.The whole authentication process is carried out by optical method without electronic calculation,so the system has a fast authentication speed.The computer simulation results prove the feasibility,security and accuracy of the optical encryption system and authentication scheme proposed in this paper.