Visually Meaningful Image Encryption Algorithm Based On Chaotic Mapping And Bayesian Compressed Sensin

With the rapid progress of remote wireless communication technology,we can now use wireless communication devices for real-time information exchange anywhere,and the form of information storage has changed from single text to multimedia data that can convey more information content.As one of the new information exchange media,digital images containing privacy information will also face huge potential security risks when transmitted over public channels.The guarantee of security issues of private image data has become a challenging problem in today’s society.Chaotic system is an inherent random phenomenon in nonlinear system that has the characteristics of unpredictable motion trajectory,sensitivity to the initial values and control parameters of the system,and ergodic motion trajectory;its characteristics are similar to the basic requirements of cryptography.The connection and structural similarity between chaotic system and cryptography have inspired people to apply chaotic system to cryptography.Compressive sensing can be used and integrated into image encryption system because of its dimensionality reduction and random projection characteristics,which can realize simultaneous compression and encryption of one or more images,making it widely used in image encryption field.Therefore,this paper designs an efficient,secure and robust image encryption algorithm by combining chaotic mapping,compressive sensing and data embedding techniques.The main work of this paper is as follows:First,a new cosine chaotic mapping based on cosine function is proposed,and it is concluded by experimental analysis that it has excellent chaotic performance.Secondly,the plaintext image coefficient matrix is compressed and encrypted by scrambling algorithm,linear measurement and bidirectional modulo diffusion.Finally,embedding operation is performed to hide the carrier image losslessly in the secret image to obtain a ciphertext image with visual significance.The experimental results show that this encryption algorithm meets various security indicators and can resist various common attacks.Aiming at the problems existing in the previous scheme,it is improved from three aspects.The new cosine chaotic mapping proposed is optimized,and a new two-dimensional new cosine chaotic mapping is proposed,which achieves better results in chaotic performance and also has good efficiency.Secondly,by using both pixel value channel and transparency channel of png image,dual image encryption is realized.And,an adaptive threshold module is introduced,which can improve the reconstruction quality of the decrypted image.The algorithm is improved in terms of security,robustness and transmission efficiency.