DETECTION OF EDITING SIGNS THROUGH FRAME SEQUENCE ANALYSIS BY EXAMINATION OF VIDEO-SOUND RECORDING

Abstract

In the modern world, where video and audio recordings have become everyday means of communication, forensic examination of video and sound recordings has become increasingly important. This is due to the fact that video and audio materials can be used as evidence in court proceedings, but their proper interpretation and evaluation require high qualifications and experience. Video editing is a common type of video processing that can be used to create many types of media content, including films, commercials, and video blogs. However, editing can also be used to create false media content with deliberately false information (disinfor-mation), for example, fake videograms - a new videograms with the necessary transfor-mation of the original content is created from previously recorded information using vari-ous types of manipulations (combining two or more parts of one or more videograms, etc.). The article describes the method of video editing through analysis of frame sequenc-ing is presented, without the use of specialized software. This method is based on analyzing the peculiarities of the formation of the sequence of video frames in a group of images in the video, including the creation feature of individ-ual types of frames. The peculiarities of three types of video frames, their formation and order of for-mation into groups in the video stream are described and analyzed. Specific examples of the order of frames used in typical codecs of Moving Picture Experts Group - H.261 and H.264 are presented. An optimal algorithm for identifying signs of editing has been proposed, which de-termines these signs based on the general principles of forming the sequence of frames in a videograms, as well as the characteristics of each type of video frame, without using spe-cialized software. Empirical examples indicate that relying solely on the detected signs of violations in the formation of frames and their sequence is insufficient. Therefore, to establish the truth-fulness of the presence of editing in the videograms, the detection of signs of editing is car-ried out in conjunction with other methods of analysis. The synthesis of all detected signs of editing using as many methods as possible is presented as a result of the research to ob-tain the most accurate result. The obtained results can be useful for the development of au-tomated systems for detecting fake videos and combating disinformation.