Image Encryption Algorithm Based On Semi-tensor Product Compressed Sensing

In the Internet of Things and big data era,multimedia data is acquired by various multimedia sensors and mobile devices every day.The collection of these multimedia big data needs to deal with two main challenges.The first challenge is that due to the limited computing resources and large data volume of sensors and mobile devices,it is necessary to implement low-cost sampling compression coding.Compression coding can reduce transmission bandwidth consumption,thereby saving the power consumption of sensors and mobile devices.The second challenge is to prevent illegal users from extracting valuable information during sampling and transmission.In order to meet the requirements of the above challenges,compressed sensing has become a promising method for data collection in the Internet of Things.Compressed sensing,as a novel signal acquisition theory,has quickly attracted a lot of attention in many fields since it was proposed.Interestingly,the compressed sensing framework can be seen as an encryption scheme.Compared with traditional encryption schemes,compressed sensing encryption schemes have many advantages such as low computational cost of encryption process,simultaneous realization of encryption and compression,and robustness of ciphertext.It can be said that the compressed sensing framework for information protection has broad application prospects.This thesis starts from the image encryption protection of compressed sensing,studies the chaos theory,semi-tensor product,information hiding and message authentication mechanisms that affect the factors of image data content and their modes of action,and constructs image encryption algorithms based on compressed sensing.On the basis of relevant research results,two encryption algorithms are proposed,and theoretical introduction and experimental analysis are carried out.The main research contents of this paper are as follows:(1)In traditional compressed sensing,the measurement matrix has been facing problems such as large data storage,high memory usage,and large data calculations.In order to reduce the storage space of the measurement matrix,improve the memory utilization,and reduce the amount of data,this paper proposes a compression sensing strategy based on semi-tensor product.The semi-tensor product is used to extract the sparse matrix from the measurement matrix,and the Arnold scrambling algorithm is used to improve the security of the secret image.In addition,in traditional image encryption,the image is transmitted in the form of ciphertext in the channel,which is easy for hackers to find and attack the encrypted image.Therefore,this paper proposes a visual security strategy for encrypted images,which converts encrypted images into visually significant encrypted images.(2)The challenge in the era of big data is how to realize low-cost data collection while ensuring data confidentiality.In this paper,we propose a message-verifiable image encryption method based on semi-tensor product compressed sensing.First,the cascade chaotic system is used to generate the measurement matrix.Then,the measurement matrix and the verification matrix are used to sample the original signal by a semi-tensor product method to obtain the measurement value matrix and identity verification matrix.The measured value matrix is further used for Arnold scrambling to obtain the final encrypted image.At the same time,the identity verification code generated by the identity verification matrix is transmitted on the public channel together,and the initial seed of the cascaded chaotic system is transmitted on the secure channel as the key.