We Need No Pixels: Video Manipulation Detection Using Stream Descriptors

TL;DR

This paper introduces a scalable method for detecting manipulated videos by analyzing their stream descriptors with simple classifiers, avoiding pixel analysis and effectively identifying forgeries unless stream data is carefully sanitized.

Contribution

It proposes a novel approach that uses multimedia stream descriptors for video manipulation detection, bypassing pixel-based analysis and demonstrating high effectiveness on standard datasets.

Findings

High detection accuracy with stream descriptor analysis

Effective against un sanitized stream data

Scalable and simple classification approach

Abstract

Manipulating video content is easier than ever. Due to the misuse potential of manipulated content, multiple detection techniques that analyze the pixel data from the videos have been proposed. However, clever manipulators should also carefully forge the metadata and auxiliary header information, which is harder to do for videos than images. In this paper, we propose to identify forged videos by analyzing their multimedia stream descriptors with simple binary classifiers, completely avoiding the pixel space. Using well-known datasets, our results show that this scalable approach can achieve a high manipulation detection score if the manipulators have not done a careful data sanitization of the multimedia stream descriptors.