МОДИФІКАЦІЯ АЛГОРИТМУ ВИЯВЛЕННЯ КЛОНУВАННЯ КАДРІВ У ВІДЕОПОСЛІДОВНОСТЯХ

The work developed a modification of the algorithm for detecting cloning of frames in video sequences. It is impossible to imagine the life of a modern person without the constant presence of digital technologies in it. They have become an integral part of our daily life. Everyone has a smartphone, almost everyone owns a personal computer or laptop. Photo and video recording technologies, as well as the distribution and publication of images and videos, such as social networks, have become an integral part of digitalization. Digitization has also spread to very important areas for human life, such as forensics or mass media. In order to have confidence in video materials, which is extremely important in criminal investigations, court cases and information networks, it is necessary to develop the field of exposing modifications and falsifications in video and photo materials, develop new methods that allow you to verify the integrity of the video or detect falsification, and modify and improve already developed ones. The paper analyzes the methods for finding falsifications in video sequences and modern formats and algorithms for saving and compressing video sequences. As a result of the analysis, the current state of development in the field of detection of fake videos was revealed and the most popular video formats were determined. Experiments were conducted with videos of different formats and resolutions, data were collected and analyzed, after which the necessary threshold values of the Pearson correlation coefficient were established for each format studied. Threshold values for searching for cloned frames were proposed in the work. The threshold value is chosen depending on the format and resolution of the video. Parameters such as video format and resolution were taken to modify the algorithm for detecting cloning of frames in a video sequence by analyzing the threshold values of the Pearson correlation coefficient. A software implementation of the modification of the algorithm for detecting cloning of frames in video sequences was created. Errors of the first and second kind are cited in the work. [ABSTRACT FROM AUTHOR]