Improvement of hybrid data compression algorithms and practical applications

The most important problems of recent years are undoubtedly the high size of data and the transmission or storage of data over insecure networks. Compression and encryption algorithms are generally used to overcome these problems. One of the approaches used for encryption today is block cipher algorithms. One of the most important cryptological components to meet the confusion requirement in block cipher algorithms is s-boxes. Because s-boxes are nonlinear structures, making the algorithm resistant to differential cryptanalysis. Therefore, the higher the nonlinearity of the s-box, the more attack resistant it will be. The most important disadvantage of chaos-based s-box structures is that their nonlinearity is lower than mathematically-based s-box structures. In this thesis, firstly, seven different approaches are proposed to increase the nonlinearity values of chaos-based s-box structures. All of these approaches have improved the performance of s-boxes. Since the JPEG algorithm is widely used on many platforms, its importance is an undeniable fact. However, improvements that can be made to meet users' needs for more data storage and faster data transmission/processing have gained momentum with the digital transformation. As quantization tables from these studies are a critical component that affects the success of the algorithm, improvement studies in this area have focused on the design of different quantization tables. In this study, a method based on chaotic systems is proposed for the first time to create different quantization tables. The proposed method has a simple structure and achieves high compression ratios. Despite achieving a high compression ratio, not sacrificing image quality is another advantage and unique aspect of the proposed method. Another approach used in the literature to improve the performance of the JPEG algorithm is to generate quantization tables using optimization methods. When the proposed method is compared with these approaches, the lower computational complexity stands out as another advantage of the proposed method. These results show that improvements can be made in future studies for many application areas such as the internet of things, industry 4.0 and artificial intelligence studies, significant storage savings and faster data transmission/processing can be achieved.