Videolarda sahtecilik tespiti

The whole world experiences digitalization at a rapid pace. Recently, smart phones, internet opportunities and digital video recording devices have accelerated digitalization. In social media, video sharing has become incredibly widespread. At the same time, videos are the most reliable and vital evidence both in social media and especially in courts. Videos are trusted more than pictures and any other evidence. However, many powerful and inexpensive video authoring tools for modifying video content are easily available by everyone today. A malicious attacker can destroy evidence in videos by deleting a sequence, adding a new sequence, changing the video stream order, etc. The easiness of these types of tampering has also reduced the trust to the videos. Therefore, videos can be accepted as reliable evidence by testing the originality of the video. In spite of the increasing variety of attacks on video content, the variety of authenticity testing methods increase day by day. In this study, a method is proposed to detect attacks such as deleting a part of the video content, adding a new section or changing the video stream order. Videos are formed by the flow of frame groups, such as pictures, in the time axis. Algorithm which is developed for our method is based on the using of correlation coefficients during the transitions of video frames. The method has been implemented with Matlab and tested on 120 videos. Test results have shown that the proposed method works quickly and successfully.