数字视频篡改痕迹的被动检测技术综述

Review of Passive Detection Techniques for Digital Video Tampering Traces

  • 摘要: 随着DEEPFAKE等人工智能方法对数字媒体数据的篡改方式越来越强大,再次掀起了全球各国政府、人民对数字媒体篡改问题的关注和焦虑。无论是传统的篡改攻击,还是智能的篡改生成,从结果上来看,都加深了人们对数字媒体真实性的质疑。数字视频是当下互联网信息传播的主流方式之一,因此对数字视频篡改痕迹的被动检测具有重大研究意义。本文针对数字视频篡改方法和痕迹被动检测问题展开综述,旨在阐述当下这一领域的发展现状和未来趋势。首先从已有案例对数字视频篡改的背景、篡改技术现状、社会危害进行多层次分析,凝炼该领域的研究现状和现存科学难题。然后,本文把被动取证检测技术分为时空域特征检测技术、编码域特征检测技术、深度特征检测技术三类,并详细分析了每类技术的特点和不足。其次,本文也对视频篡改检测相关的国际竞赛和国内竞赛、视频篡改数据集以及常用第三方工具库进行了归纳和总结。最后,本文总结了该领域的进一步有待研究的理论、技术内容和未来应用发展趋势。

     

    Abstract: As DEEPFAKE and other artificial intelligence methods have become more and more powerful in tampering with digital multimedia, once again, governments and people around the world are concerned about the issue of digital media tampering. Both traditional tampering attack and intelligent tampering have raised people's doubts about the authenticity of digital media. Therefore, passive detection of digital video tampering traces is of great significance. This paper gives an overview of this problem to analyze the current development status and future trends in this field. First, the background and negative impacts of video tampering are analyzed. The meaning of video tampering detection from different angles, and the difficulties of the problem are also summarized. Next, this paper analyzes different types of digital video re-editing technologies, and divides the existing passive forensic detection technologies into detection method based on spatio-temporal domain feature, coding domain feature, and depth feature, features and deficiencies are detailed in this paper. In addition, this paper also summarizes the existing international and domestic competitions, video tampering data sets, and third-party tool libraries commonly used in the field, and also points out the follow-up research content and development trends in this field.

     

/

返回文章
返回