Detecting and Locating Passive Video Forgery Based on Low Computational Complexity Third-Order Tensor Representation

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 are available and 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. In this paper, a newly developed passive video forgery scheme is introduced and discussed. The developed scheme 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 tensor data are orthogonally transformed to achieve more data reductions and to provide good features to trace forgery along the whole video. Experimental results and comparisons show the superiority of the proposed scheme 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 which are rarely tested by previous works. Moreover, the proposed scheme offers a reduction in time and a linear computational complexity. Based on the used computer’s configurations, an average time of 35 s. is needed to detect and locate 40 forged frames out of 300 frames.

Frame Duplication and Insertion Forgery Detection in Surveillance Videos Using Optical Flow and Texture Features

“…Alsakar et al [22] focused on the analysis and identi cation of frame insertion and deletion forgery in videos based on low computational complexity third-order tensor representation. Precision, Recall and F1 for detection of frame insertion forgery by the state of the art are 96%, 94% and 95% respectively.…”

“…Although various solutions to frame duplication and insertion detection have been proposed, they still face three main challenges. The rst one is limited applicability; several forgery detection techniques have restrictions on videos like video format, the number of tampered frames, and frame rate, which limit their practical applicability [22,23]. For example the method in [24] cannot detect frame duplication of more than 20 frames, and method in [25] can detect insertion and duplication if inserted/duplicated frames are multiple…”

“…In addition, these datasets are small in size. Mostly cross-validation has not been performed to ensure the generalization capability of existing methods [11,22] due to unavailability of datasets. The third one is forgery localization; where the method in [28] detects a deletion forgery in the center (between frames 8 and 9) of a single video shot.…”

“…Evaluation of proposed models for frame duplication forgery detectionTo evaluate the performance of the proposed method, we have used 80% of our dataset as a training set and 20% as a testing set.The proposed method has been compared with similar state-of-the-art methods[22] based on Harris, Gray Level Co-occurrence Matrix (GLCM), and Singular Value Decomposition (SVD) features. They used a dataset of 18 video clips having the same frame rate and same le format, taken from TRACE to conduct experiments.…”

Image and Video Forensics

“…Rodriguez-Ortega et al [1] presented a copy-move forgery detection technique based on a deep learning model to overcome the problem of generalization among different datasets. Alsakar et al [2] focused instead on the analysis and identification of forgery in videos based on low computational complexity third-order tensor representation. Two types of forgery have been considered: insertion and deletion for static and dynamic videos.…”