基于改进局部二值模式的YASS检测方法

Steganalysis Method of YASS based on Improved Local Binary Patterns

如今互联网已经逐渐渗透到人们生活的诸多方面,成为日常通信的重要途径。信息隐藏作为一种通信技术,通过将秘密信息嵌入常见载体中以达到隐蔽通信的目的。图像隐写是利用图像作为载体进行信息隐藏的一门技术与科学,YASS(Yet Another SteganographicScheme that Resists Blind Steganalysis)通过随机选取图像的子块进行DCT变换和QIM信息嵌入,具有较高的安全性。文中通过引入图像的局部二值模式(LBP)这一概念,根据YASS算法特点,分析图像的局部纹理变化,改进局部二值模式,利用局部有序对比模式(LOCP)的特征进行隐写分析。通过大量实验表明,相比传统的YASS隐写分析,文中所提方法在分析检测正确率等方面都有更好的效果。

Today the Internet gradually penetrates into every aspect of people＇s lives and becomes the important channel of daily communication. As a communication technology, information hiding embeds secret information into common hidden carriers in order to achieve concealed communication. Image steganography aims to use the image as a carrier for information hiding. YASS（yet another steganographic scheme that resists blind steganalysis） could achieve a high security level by randomly using an image sub-block DCT transform and QIM information embedding. This paper introduces the Local Binary Pattern（LBP） concept. In addition, according to the characteristics of YASS, the Local Ordinal Contrast Pattern（LOCP） feature is proposed for steganalysis upon YASS which describes the local texture variation of the image. Experimental result indicates the proposed method could achieve better performance in detecting YASS than the traditional approaches.