Logo image
Back
A new approach to detecting frame deletion in H.264 encoded digital video
Thesis   Open access

A new approach to detecting frame deletion in H.264 encoded digital video

Hunter Kippen
Master of Science (M.S.), Drexel University
May 2019
DOI:
https://doi.org/10.17918/ytkv-3a24
pdf
Kippen_Hunter_20197.15 MBDownloadView
PDF Open Access Open Access (License Unspecified)

Abstract

Electrical engineering Forensic sciences--Data processing Signal processing--Digital techniques Computer Engineering
Due to the increasing prevalence of falsified video content, it is important to be able to identify altered video recordings. For certain types of video, removing small portions of content can change the perception of the events that transpire without leaving visually detectable evidence. In particular, removing a set or sequence of frames from the beginning of a video can change the context of surrounding events, especially in newsworthy content. Digital forensic techniques have been developed to detect this type of video manipulation, but were originally proposed for use on MPEG-2 video. Advances in video encoding, particularly in MPEG-4 and H.264 have called the effectiveness of these techniques into question. In this thesis, we examine the performance of detectors built using the methodology from MPEG-2 when applied to H.264 video, and show that indeed new methods are needed to detect frame deletion in modern video. We propose a new method for extracting fingerprints left behind by frame deletion in H.264 as well as an expanded frame deletion detection framework for ensuring robustness across videos captured by different camera models as well as variation in scene content or motion. We evaluate the performance of our proposed techniques as well as examine current limitations.

Metrics

52 File views/ downloads
59 Record Views

Details

Logo image