Indian Journal of Science and Technology

Abstract

Objectives: This paper summarizes all the proposed techniques involved in digital video inter-frame forgery detection for MPEG-1, 2, 4 and H.264/AVC encoded videos. Methods/Statistical Analysis: Double compression detection techniques are classified here on the basis of footprints analyzed during detection. The detection methods designed for videos that use fixed GOP structure for first and any of the subsequent compression are different from the videos that use different GOP structure. Video inter-frame forgery detection techniques are then analyzed on the basis of type of forgery they detect and the type of codec used for video encoding. Findings: Digital videos often provide forensic evidence in legal, medical and surveillance applications but are more prone to inter-frame forgeries, which are not only easy to perform but are equally difficult to detect as well. The analysis of the literature ascertained that majority of the proposed techniques are dependent on the number of frames tampered and video codec used to encode the videos. Among these proposed techniques, double MPEG compression was best detected using the technique which utilizes Benford’s law and was proposed by Chen and Shi on MPEG-1 and MPEG-2 encoded videos. On the other hand, Wang et al. gave sound results for all kinds of inter-frame forgeries on MPEG-2 encoded videos by utilizing the measure of optical flow consistency. Since very few authors focussed on forgery detection in MPEG-4 encoded videos and thereby such techniques have not been discussed in many survey papers. Moreover, digital cameras especially surveillance cameras which generate massive amount of videos these days have built-in MPEG-4 codec because it offers a better compression rate. Application/Improvements: Video forensics domain, therefore, is in dire need of a technique that will detect any kind of inter-frame forgery in MPEG-4 encoded videos.

Keywords: Digital Forensics, Inter-frame Forgery, Video Forgery Detection, Video Forgery Detection Survey, Video Tampering Detection