基于Arnold变换的数字图像自适应隐写算法

以往的图像自适应隐写算法大多数以顺序满嵌的方式嵌入秘密信息，这类算法秘密信息的隐蔽性不够高，因此提出一种随机非满嵌算法。通过分析出图像的系统参数使图像的满嵌容量刚好大于秘密信息的长度让载体图像达到非满嵌，增强了隐写的灵活性，减少了载体图像的嵌入修改量。再使用Arnold变换对数据的嵌入顺序进行置乱，防止攻击者按顺序分析出秘密信息，使秘密信息的隐蔽性变高，进而提高了算法的安全性。实验结果表明本算法提高了隐写的隐蔽性，减少了图像的嵌入失真度，且随机非满嵌操作适用于很多同类算法。

Adaptive steganography algorithm in digital image based on Arnold transform

Previous adaptive steganography algorithms were based on ordinal and full-embedding method. However, these algorithms might cause the concealment of secret information not good enough. Therefore, a random and non-fully embedding method was proposed to improve the concealment. The image system parameters could be obtained, so that the image full capacity was exactly greater than the length of information for non-full embedding. Moreover, these parameters made steganography more flexibility by reducing the modifications of image. Then, secret information embedding sequence was scrambled based on Arnold transform for the purpose that the steganalyst cannot detect out information orderly. Therefore, the secret information can be more concealed, which can improve the security of the algorithms. Experimental results show that the algorithm not only improves the concealment of steganography but also reduces stego distortion, and random non-full embedding operating can be applied to many other similar algorithms.