Frame Shuffling Forgery Detection Method for MPEG-Coded Video

Frame duplication and frame shuffling represent prevalent methods employed by forgers to manipulate digital video content, aiming to conceal or emphasize specific activities within the video. While several forensic techniques have been developed to address frame duplication forgery, frame shuffling forgery has received limited attention, resulting in a gap in detection methods. This paper introduces a novel forensic framework integrating motion vector analysis and scale-invariant transform features to effectively detect and locate frame shuffling forgeries in moving picture expert group (MPEG)-coded videos. The proposed algorithm’s performance is rigorously assessed using precision rate, recall rate, and detection accuracy metrics across 15 test videos. Remarkably, the algorithm achieves an average detection accuracy of 100%, signifying its exceptional precision in identifying forgeries. Additionally, the algorithm demonstrates efficiency, with an average simulation time of only 36 s, underscoring its effectiveness in real-time forensic applications.