北京邮电大学学报

  • EI核心期刊

北京邮电大学学报 ›› 2006, Vol. 29 ›› Issue (2): 1-4.doi: 10.13190/jbupt.200602.1.jiangn

• 论文 •    下一篇

新的唯秘密载体信息隐藏分析方法

姜楠,王健,杨义先   

  1. 北京邮电大学 信息安全中心
  • 出版日期:2006-04-28 发布日期:2006-04-28

A New Method for Blind Image Steganalysis

JIANG Nan,WANG Jian,YANG Yixian   

  1. Information Security Center, Beijing University of Posts and Telecommunications
  • Online:2006-04-28 Published:2006-04-28

摘要:

将图像质量度量标准进行改造并结合在一起,可以针对图像数据进行唯秘密载体(盲)信息隐藏分析,即将图像质量度量标准中的原始图像改为全白图像设计特征向量,用 SVM(支持向量机)对特征向量分类。分析过程中引入的人眼视觉系统的带通滤波性质,提高了算法准确率。实验结果表明了该方法对分析自然图像是有效的。

关键词: 信息隐藏, 信息隐藏分析, 人类视觉系统, 支持向量机

Abstract: Image quality metrics have two inputs: the original image and the stego image. Whereas blind image steganalysis only has one input: the stego image. In order to fit blind image steganalysis, the image quality metrics must be altered: the original images are replaced by white images. Using four image quality metrics, the distance between stego images and white images can be obtained. Regarding the distance as eigenvectors and using SVM (support vector machine) to classify them into two classes: one includes hidden information, the other one doesn't. Bandpass filter property of human visual system is used to improve correct detection rate. Experiments show that this method is effective to nature images.

Key words: Information hiding, steganalysis, human visual system, support vector machine