Chaos-based image encryption in internet of things applications

With the development of information systems and the proliferation of Internet of Things (IoT) devices, data security has gained critical importance. Unlike standard computer systems where complex algorithms can be used, the limited hardware capacities of IoT devices necessitate lightweight and customized encryption solutions. With this vision, in this study, two different lightweight and efficient image encryption methods are proposed in the category of chaos-based image encryption algorithms for IoT environments. The first method, Corner Transition Algorithm (CGA), is in the category of image encryption by pixel mixing method. Integrated with chaos-based propagation methods, KGA creates a comprehensive encryption scheme and exhibits stringent statistical and differential properties. In the KGA algorithm, studies with pixel mixing algorithms were compared and the advantages of our proposed algorithm were tried to be revealed. The processing times of KGA and three similar algorithms were compared. As a result, in all tests performed, the proposed algorithm gave better results than its alternatives. In addition, comparisons between relevant studies were evaluated based on UACI, NPCR, PSNR and entropy values. As a result, the best results for UACI and NPCR were obtained from the proposed algorithm. Other evaluation metrics, PSNR and entropy, are also within the accepted range. The second xviii method offers a multi-layer Nonlinear Permutation Framework (DOPF) 2D matrix transformation based encryption approach. Unlike standard 2D matrix mapping methods, the number of steps on the permutation matrix of DOPÇ is variable and the encryption complexity is increased by using 3 different 1D chaotic maps (Zaslavsky, Chebyshev and logistic maps) in the implementation stages of the algorithm. Detailed performance evaluations of both methods are presented and compared with similar methods. Test results show that the proposed algorithms achieve high encryption complexity and low processing time compared to alternatives. The KGA method meets the stringent security requirements of IoT systems, showing little deviation from the expected value with values of 99.6093 and 33.4648 in the NPCR and UACI evaluation criteria. The DOPÇ method offers high efficiency by completing the encryption steps in a single image navigation. In the tests performed on 14 SIPI dataset images, DOPÇ average NPCR values of 99.6122, UACI of 33.4690 and information entropy of 7.9993 were obtained. Both methods have been shown to be resistant to brute force attacks thanks to their high key space.