在基于失真不可感知的控制模型(maximize payload under imperceptive,MPUI)下,研究了空域图像的最大水印容量。发现最大水印容量不仅与嵌入强度有关,而且与图像尺寸大小、复杂度、视觉敏感度等因素相关。在理论上推导了最大水印容量与图像大小、嵌入强度的关系,并采用多元回归分析方法研究了图像复杂度、视觉敏感度与嵌入率的关系模型。综合这些内外因素,借助于主客观评价体系,推算出水印嵌入容量的估算方法,并在实验中验证了该估算方法的有效性。
In order to estimate maximum steganographic capacity of discrete cosine transform(DCT) domain in JPEG image, this paper presents a method based on the maximize capacity under undetectable model(MCUU). We analyze the relation between steganographic capacity and affecting factors(image size, steganography operator, loading band, embedding intensity and image complexity). Then we design a steganography analyzer architecture and a steganographic algorithm which can dynamically increase the steganographic capacity. Compared with other methods of embedding capacity estimation in DCT domain, the proposed methods utilizes general steganalysis methods rather than one specific steganalysis technique and takes five essential factors into account, which improves the commonality and comprehensiveness of capacity estimation, respectively. The experimental results show that steganographic capacity for quantization index modulation(QIM) is almost twice that of spread spectrum(SS) based on MCUU model.
MAO JiafaHUANG YanhongNIU XinxinXIAO GangZHU LinanSHENG Weiguo
