是否存在一种可以抵抗图像操纵的数字图像隐写算法？
我想知道 - 是否有一种针对数字图像的隐写解决方案可以抵抗图像操作？我所说的“操作”是指最标准的操作 - 重新压缩 JPEG（甚至完全更改文件格式）、裁剪和缩放。这种方法的应用自然是为了图像版权的保护。
我完全理解，图像被操纵的越多，隐写水印完好无损的机会就越少，但至少有一定程度的抵抗力会很好。
I'm wondering - is there a steganography solution for digital images which is resistant to image manipulations? With "manipulations" I mean the most standard operations - recompressing JPEGs (or even changing file formats entirely), cropping and scaling. The application of this method would naturally be for image copyright protection.
I fully understand that the more image is manipulated the less are the chances that the steganographic watermark is intact, but at least some degree of resistance would be nice.
如果你对这篇内容有疑问，欢迎到本站社区发帖提问 参与讨论，获取更多帮助，或者扫码二维码加入 Web 技术交流群。
绑定邮箱获取回复消息
由于您还没有绑定你的真实邮箱，如果其他用户或者作者回复了您的评论，将不能在第一时间通知您！
发布评论
评论（1）
在我看来，这只是猫捉老鼠游戏中的又一举动，但你在这里：
http://scien.stanford.edu/class/ee368/projects2000/project5/default.htm
从结论中：
“水印嵌入总是存在权衡。通过该权衡，我们水印的鲁棒性与图像的退化之间的平均权衡 在某些时候，水印的强度将导致图像退化到一个点，这取决于
用于水印的技术，无论是在空间上 。或频域，加水印的图像总体上表现出对同一域的攻击的鲁棒性和对相反域的攻击的弱点，Nakamura 在他的工作中提到了这些结果，并得出结论，没有任何方法在该域中是鲁棒的。这两个领域，并且应用混合水印方案将会产生更糟糕的结果。”
It seems to me, that this is just another move in the cat and mouse game, but here you are:
http://scien.stanford.edu/class/ee368/projects2000/project5/default.htm
From the conclusions:
"There will always be a tradeoff in watermark embedding. By which we mean tradeoff between robustness of the watermark verses the degradation of the image. At some point the strength of the watermark will cause the image to be degraded to a point, which will not be acceptable.
Depending on the technique used for watermarking, whether in spatial or frequency domain, the watermarked image showed, in general, robustness to the attacks of the same domain and weakness to the attacks in the opposite domain. Nakamura mentions these results in his work, and concludes that there is no method that will be robust in both domains, and that applying hybrid watermarking schemes will have worse results."