| With the rapid development of the Internet and social networking platforms,digital images are gradually and widely transmitted on the network as a mainstream sharing tool.The information present in the images may involve personal privacy or even more confidential content,which may often be illegally intercepted,tampered with,or even destroyed.Therefore,encryption protection of images is a necessary tool used to safeguard information security.Chaos is a complex physical phenomenon with high pseudo-randomness,strong sensitivity to initial values and long-term unpredictability,which are very suitable for encryption of digital images.This paper constructed two new types of chaos models,and on this basis,image encryption schemes are designed by using the improved chaos systems.In addition,considering the actual application environment of image encryption algorithms,the two types of image encryption algorithms proposed in this paper encrypt the whole as well as the local images to achieve image privacy protection in different application scenarios.Firstly,this paper constructs a one-dimensional quadratic chaotic map based on topological conjugation theory and proves theoretically that the proposed onedimensional quadratic map is chaotic by designing a homeomorphism to the Logistic chaotic map.Compared with the disadvantage that the Logistic map has only one adjustable parameter,this map has multiple adjustable parameters,which ensures its significant scalability in the parameter space than the traditional one-dimensional chaotic maps without increasing the cost,which proves that the proposed onedimensional quadratic chaotic map is more suitable for the design of cryptographic algorithms.Based on this,this paper designs an encryption algorithm to encrypt the entire images,which performs row-by-row and column-by-column transformation of pixels according to the generated chaotic sequence.In the diffusion algorithm,the pixel values are rearranged by bit using a one-dimensional quadratic chaotic sequence,and the bit xor operation is used to vary the pixel values.After experiments,it is proved that the algorithm has superior security performance and is competitive compared to other chaotic image encryption algorithms.Secondly,a novel LS chaotic system combining Sine and Logistic map is proposed to improve the nonlinear dynamics of this chaotic system mainly by perturbing its control parameters and introducing delay states.We obtain portrait content in the image using the PSPNet semantic segmentation model trained on the Cityscapes dataset and the Faster-RCNN target detection model trained on the MSCOCO dataset,and an image encryption algorithm is designed for this partial image content.In this algorithm,the chaotic sequences are sorted and then the pixel values in the partial image are gradually replaced using their index arrays.After that,the first 4 bits of the image pixel values are diffused using the index sequence as well as the generated chaotic sequence.After performance testing and analysis experiments,the algorithm has strong resistance to statistical attacks and shows superior security performance compared with other encryption algorithms. |
