合成图像取证是数字图像取证的一个重要研究内容,研究合成图像被动取证技术对确保数字图像的原始性和真实性有着极其重要的意义.根据合成图像被动取证技术研究角度不同.将现有的技术方法分为3大类:基于传统特征的合成图像取证、基于盲源分离的合成图像取证、基于稀疏表示的合成图像取证.然后分别介绍了3类取证技术的基本特征和典型算法.并对不同方法进行了分析比较和总结.最后指出当前研究中存在的一些主要问题.给出本领域未来的研究方向.
Composite image forensics is a vital research area of digital image forensics. The research on techniques of composite image passive forensics is important to ensure the originality and authenticity of digital image. According to different research areas, the techniques of composite image passive forensics fall into three categories: techniques based on traditional characteristics, techniques based on blind source separation, and techniques based on sparse representation. The basic characteristics, typical methods, as well as comparison and analysis of various algorithms are summarized in detail for each category. The main problems in the current research are pointed out and the future directions are presented.