An adaptive LSB matching steganography based on octonary complexity measure
An adaptive LSB matching steganography based on octonary complexity measure
复制标题
DOI：
10.1007/s11042-011-0975-y
复制
发表时间：
2012-01
影响因子：
3.6
通讯作者：
V. Sabeti;S. Samavi;S. Shirani
中科院分区：
文献类型：
--
作者：
V. Sabeti;S. Samavi;S. Shirani
自适应隐写技术可以提高隐写的安全性。大多数自适应方法使用LSB翻转（LSB-F）来嵌入部分算法。LSB-F对简单的隐写分析方法非常脆弱，但它允许自适应算法在接收器侧可提取。使用LSB匹配（LSB-M）可以提高安全性，但在接收器处提取数据很困难，有时甚至是不可能的。有许多针对LSB-M的攻击。在本文中，我们提出了一种自适应算法，与大多数自适应方法不同，使用LSB-M作为其嵌入方法。所提出的方法使用了一个复杂的措施的基础上确定的安全位置的图像的局部邻域分析。使用LSB-M的可比自适应方法在执行嵌入时遭受像素复杂度的可能变化。所提出的算法是这样的，当一个像素被归类为复杂的发射机和嵌入的接收器将识别它作为复杂的，数据被正确地检索。通过获得更高的PSNR值的嵌入图像相对于可比的自适应算法的算法的更好的性能示出。该算法对多种攻击的安全性高于LSB-M。此外，它与最近的自适应方法进行了比较，并被证明是有利的大多数嵌入率。
Adaptive steganography methods tend to increase the security against attacks. Most of adaptive methods use LSB flipping (LSB-F) for embedding part of their algorithms. LSB-F is very much vulnerable against simple steganalysis methods but it allows the adaptive algorithms to be extractable at the receiver side. Use of LSB matching (LSB-M) could increase the security but extraction of data at the receiver is difficult or, in occasions, impossible. There are numerous attacks against LSB-M. In this paper we are proposing an adaptive algorithm which, unlike most adaptive methods, uses LSB-M as its embedding method. The proposed method uses a complexity measure based on a local neighborhood analysis for determination of secure locations of an image. Comparable adaptive methods that use LSB-M suffer from possible changes in the complexity of pixels when embedding is performed. The proposed algorithm is such that when a pixel is categorized as complex at the transmitter and is embedded the receiver will identify it as complex too, and data is correctly retrieved. Better performance of the algorithm is shown by obtaining higher PSNR values for the embedded images with respect to comparable adaptive algorithms. The security of the algorithm against numerous attacks is shown to be higher than LSB-M. Also, it is compared with a recent adaptive method and is proved to be advantageous for most embedding rates.