基于遗留痕迹的数字图像盲取证综述

Abstract

数字图像盲取证是一种新颖的图像信息安全技术，它能够在没有前期预签名或嵌入信息等先验条件下，完成对图像真伪和来源的认证，具有更广的实用价值，成为目前安全领域新的研究热点。本文以基于遗留痕迹的数字图像盲取证为代表，从复制–粘贴检测、模糊篡改检测、JPEG相关检测、照明不一致检测和重采样检测五个方面进行综述，对与之相应的典型算法进行了重点介绍，并对未来的发展进行展望。 Blind digital image forensics is a new kind of technique to protect the security of image information. It can detect image authenticity and source without any pre-extraction or pre-embedded information. With great practical value, it has been a new research focus of the current field of security. Taking Blind digital image forensics which is based on the traces of tamper for an example, this paper makes a summary from five aspects: copy-paste forgery, blurring detection, JEPG detection, inconsistence detection in the light source direction and resampling. The typical algorithms are introduced selectively and the future research and development tendency are pointed out.