【作者】 刘绍辉； 龚龙； 姚鸿勋； 高文；

【Author】 Liu Shaohui, Gong Long, Yao Hongxun, Gao Wen School of Computer Science and Technology, Harbin Institute of Technology, Harbin, 150001, China Institute of Computing Technology, the Chinese Academy of Sciences, Beijing 100080, China

【机构】 哈尔滨工业大学计算机学院；

【摘要】 隐写分析技术是指发现信息隐藏的事实,使其失效,并对隐藏算法参数作进一步估计的技术。针对最简单的空域LSB隐写技术,其隐写分析技术已经比较成熟,但是对于自适应隐藏技术,其隐写分析技术则很少。分析从预测差分图像出发,提出了改进的RS(Regular Singular)隐写分析算法,在此基础上构造了能有效区分隐写图像和原始图像的统计量,并依次设计了判断隐藏信息量的公式和隐藏区域的算法。实验结果表明,算法对于采用自适应空域隐藏算法进行隐藏的图像,不仅能有效的判断图像是否隐藏有信息,而且能准确的计算嵌入的信息量和嵌入的位置。更多还原

【Abstract】 Information hiding deals with hiding information into cover signals, and it must modify some statistical features of the original signals without affecting the perceptual quality. Steganalysis is to break information hiding, including identifying the type of algorithm, estimating information size and embedded region and so on. Many existing steganalysis algorithms aimed to simple LSB hiding methods but not to adaptive hiding techniques. This paper proposed a novel steganalysis algorithm based on predicted difference image to adaptive hiding techniques, and then constructed statistics to distinguish original images from steg images, estimating the size of the information hidden into original image and locating the area with hidden information. Extensive experiments indicate this algorithm not only can determine whether image has been hidden information or not but also can estimate accurately the size of information and embedded area.更多还原