网络与信息安全学报 ›› 2021, Vol. 7 ›› Issue (5): 13-28.doi: 10.11959/j.issn.2096-109x.2021047

• 专栏Ⅰ:语音图像与音视频处理 • 上一篇    下一篇

基于深度学习的数字图像取证技术研究进展

乔通1, 姚宏伟1, 潘彬民1, 徐明1, 陈艳利2   

  1. 1 杭州电子科技大学网络空间安全学院,浙江 杭州 310018
    2 法国特鲁瓦工程技术大学,特鲁瓦 10000
  • 修回日期:2020-09-25 出版日期:2021-10-15 发布日期:2021-10-01
  • 作者简介:乔通(1986− ),男,河南新乡人,博士,杭州电子科技大学副教授,主要研究方向为数字图像取证、信息隐藏
    姚宏伟(1993− ),男,福建泉州人,杭州电子科技大学硕士生,主要研究方向为数字图像取证、人工智能安全
    潘彬民(2000− ),男,浙江温州人,主要研究方向为图像隐写分析
    徐明(1972− ),男,浙江杭州人,博士,杭州电子科技大学教授、博士生导师。主要研究方向为数字图像取证、网络安全
    陈艳利(1992− ),女,山东青岛人,法国特鲁瓦工程技术大学博士生,主要研究方向为数字图像取证
  • 基金资助:
    浙江省属高校基本科研业务费专项资金(GK219909299001-007);国家自然科学基金(61702150);浙江省基础公益研究项目(LGG19F020015);网络空间安全重点专项基金(2016YFB0800201)

Research progress of digital image forensic techniques based on deep learning

Tong QIAO1, Hongwei YAO1, Binmin PAN1, Ming XU1, Yanli CHEN2   

  1. 1 School of Cyberspace, Hangzhou Dianzi University, Hangzhou 310018, China
    2 University of Technology of Troyes, Troyes 10000, France
  • Revised:2020-09-25 Online:2021-10-15 Published:2021-10-01
  • Supported by:
    The Fundamental Research Funds for the Provincial Universities of Zhejiang(GK219909299001-007);The National Natural Science Foundation of China(61702150);The Public Research Project of Zhejiang Province, China(LGG19F020015);The Cyberspace Security Major Program in National Key Research and Development Plan of China(2016YFB0800201)

摘要:

随着数字图像篡改技术不断的革新换代,传统的取证方法已经无法对抗最新的多媒体篡改手段和技术,尤其是深度造假及深度学习技术带来的全新挑战。总结提炼了包括图像预处理模块、特征提取模块及分类结果后处理模块的通用数字图像取证框架,并在提出的框架基础之上分析现有相关研究的优缺点,同时归纳了数字图像取证面临的挑战并指明未来的发展方向。

关键词: 数字图像取证, 卷积神经网络, 来源识别, 篡改检测

Abstract:

In the new era of rapid development of internet, where massive forgery images with updated tampering techniques flood into, traditional algorithms are no longer able to deal with the latest multimedia tampering techniques, especially those caused by Deepfake and deep learning techniques.Thus, a universal framework for image forensics including image pre-processing module, feature extraction module and post-processing module designed for specific classification were proposed creatively.Accordingly, the state-of-the-art algorithms were reviewed,and meanwhile the main strength and limitations of current algorithms were generalized.More importantly, the future studies were also listed for advancing the development of digital image forensics.

Key words: digital image forensic, convolution neural network, origin identification, forgery detection

中图分类号: 

No Suggested Reading articles found!