الفهرس 
Table of ContentsOnly 14 pages are availabe for public view
 |  | from | 120 |  |  |
| | | from | 120 | | |
Abstract
Great attention is paid to detecting video forgeries nowadays, especially with the widespread sharing of videos over social media and websites. Many video editing software programs such as Adobe Video Editor, Photoshop, Premiere by Adobe, and Windows Movie Maker are available. They perform well in tampering with video contents or even creating fake videos. Forgery affects video integrity and authenticity and has serious implications. For example, digital videos for security and surveillance purposes are used as evidence in courts. Video forensic deals with comparison, examination, and analysis of videos for detecting and locating forgery or tampering. Video forensics is concerned with video source identification, authentication, and video tampering detection problems. The video forensic techniques may be classified into active and passive forensic methods. Information pre-computation and embedding techniques, in which watermarks or digital signatures are inserted into videos, are examples of active forensic approaches. However, hardware chips and embedded software are required in active techniques. These methods fail when tampering occurs before adding digital signatures or watermarks. To overcome these problems, passive forensic methods detect the authenticity of given videos. This thesis targets enhancing passive methods to detect and identify forgery in given videos. In this thesis, a newly developed passive video forgery technique is introduced and discussed. The developed technique is based on representing highly correlated video data with a low computational complexity third-order tensor tube-fiber mode. An arbitrary number of core tensors is selected to detect and locate two serious types of forgeries which are: insertion and deletion. These tensors provide the natural way to represent data such as images and video. These tensor data are then orthogonally transformed to achieve more data reductions and to provide good features to trace forgery along the whole video. Two techniques for detecting inter-frame passive forgeries are developed. The proposed framework of these two techniques consists of three stages, which are 3D-tensor decomposition, forgeries detecting, and forgeries locating. The first technique for detecting the forgeries is based on correlation between consecutive third-order tensors features. Identification the forgery in video is based on a threshold. The second is based on fuzzy logic to provide in memory storage. For locating these forgeries such as insertion and deletion cases, the correlation is used and based on threshold the position of forgery in given videos is detected. Experimental results and comparisons show the superiority of the first proposed technique with a precision value of up to 99% in detecting and locating both types of attacks for static as well as dynamic videos, quick-moving foreground items (single or multiple), zooming in and zooming out datasets. They are rarely tested by previous works and these results in case of using correlation or fuzzy logic in detecting stage. The second proposed technique using fuzzy logic achieves the superiority with a precision value up to 99.4%. Moreover, the proposed technique offers a reduction in time and a linear computational complexity. Based on the used computer’s configurations, an average time of 65 s. is needed to detect and locate forged frames in the given videos.