نبذة مختصرة : One of the most common representations of information today are digital images, unauthorized changes of which can lead to negative consequences for an individual, institution, firm, and the state as a whole, which makes the detection of digital image integrity violation one of the most urgent tasks of information security. The main drawback of the existing expert methods is their focus on detecting the results of a specific disruptive action, but in practice the expert often does not have information about the specifics of an attack on a digital image, while his set of tools is always limited, which can lead to a situation where the investigated digital image is erroneously recognized as original. The first «defense line» here should be methods that are effective regardless of the type of disruptive action, i.e. universal. At present, there are a very small number of such methods in open sources; most of them are not free from shortcomings, the main of which is a significant decrease in efficiency in conditions of minor disturbances. The aim of the paper is to develop a theoretical basis for an effective universal method of detection of digital image integrity violation, in particular, in conditions of minor disturbances. In the course of achieving the aim of the paper: the justified expediency of using a block approach when organizing an examination of the integrity of a digital image; the area of discrete cosine transformation (DCP) of the block is chosen as the area of examination; justified selection of specific DCP coefficients for organizing the detection of violations of the integrity of digital images, the values of which do not depend on the value of the quality coefficient used when obtaining the original image, as well as on the specific type of digital images; the difference in the nature of changes in the selected formal parameters during re-saving of lossy digital images, depending on whether it is original or non-original, is investigated. The obtained results of theoretical studies, which are ...
No Comments.