A Universal Method of Digital Content Integrity Expertise

In today’s digital age, the majority of information utilized across various fields is visual in nature, often presented as sequences of digital images and videos. This paper considers such material as digital content, emphasizing the growing importance of safeguarding it from unauthorized access. The challenge of detecting unauthorized changes in digital content remains unresolved due to its complexity, inadequate theoretical foundation, evolving capabilities of various graphic editors, and the emergence of new steganography methods. This study aims to enhance the efficiency of digital content integrity violation detection by developing a novel, universal expert method for analyzing digital video and image sequences. The objective is achieved by developing a comprehensive analysis approach for information systems based on perturbation theory. This approach involves investigating the properties of the trend of the function of dependence of singular numbers perturbations of the image matrix/video frame on its number. It has been demonstrated that the trend behavior differs for original and non-original content. The qualitative and quantitative characteristics of these differences have been identified, and on their basis, a method of digital video integrity examination and its algorithmic implementation have been developed. These are the primary results of the work. The significance of the obtained results lies in the following: increasing the efficiency of detecting digital content integrity violations in conditions of small perturbing impacts (in the case of steganalysis, in conditions of 0.1 bpp bandwidth of the hidden communication channel); ensuring high absolute efficiency and universality of the developed method under arbitrary impacts.