以往的自适应JPEG隐写算法大多只在秘密信息嵌入之前计算图像失真,无法在嵌入过程中动态地调节失真代价.考虑到信息嵌入时的交互影响,提出一种动态更新失真代价的自适应JPEG隐写算法.首先,分析量化步长、待嵌DCT(discrete cosine transform)系数绝对值以及扰动误差3种影响嵌入波动的因素.进而,提出失真代价更新策略EUS(embedding update strategy),使图像的失真代价能够动态地更新.最后,结合此策略设计并实现一种自适应JPEG隐写算法.实验表明该算法能够显著提高JPEG隐写的安全性.%Previous adaptive JPEG steganography algorithms mostly calculate image distortion before secret informa-tion embedding, so they can't dynamically adjust distortion costs. Considering the mutual impacts during embedding, an adaptive JPEG steganography algorithm based on distortion cost updating was proposed. Firstly, three factors that affect embedding fluctuations were analyzed, including quantization step, the absolute value of the quantized DCT coefficient to be modified and perturbation error. Then the embedding update strategy (EUS) was proposed, which enabled to dy-namically update the distortion costs. After that, an adaptive JPEG steganography algorithm was implemented combining the strategy. The experimental result illustrates that the algorithm can significantly improve the secure performance of JPEG steganography algorithm.
展开▼
