On digital video watermarking using unequal error protection and

图像失真与数字水印信道容量的关系研究通过对数字水印模型的简化，从理论上研究了去除数字水印的难易程度、嵌入信息量与图像失真之间的关系。因此，我们得到了数字水印的信道容量（攻击后剩余信息的数量）与图像失真之间的关系。提出了一种不对称数字水印的思想，其中嵌入算法与提取算法不同。这就保证了即使提取方法对公众开放，也很难去除数字水印。提出了一种具体的方法，并对其抗攻击能力进行了评价。本文提出了一种适用于MPEG电影的数字水印方法，该方法在提取阶段不需要原始图像。它不需要在嵌入过程中进行解码/重新编码，因此嵌入水印的成本小。我们评估了它对实际攻击的鲁棒性，如重新编码攻击和信号处理攻击。将拼接技术与水印相结合，增强了对旋转等几何攻击的鲁棒性。针对恶意提供者的安全内容分发协议和安全叛徒跟踪方案研究提出了一种可以跟踪泄露分布式信息的叛徒的安全信息分发协议。然后，我们提出将其应用于安全内容分发协议和安全叛逆者跟踪方案中，以防止恶意提供者的攻击。

Study on the relation between image distortion and channel capacity of digital watermarkingBy simplifying the model of the digital watermarking, we theoretically investigated the relation among the difficulty of removing a digital watermark, the amount of embedded information and the distortion of the image. As a consequence, we obtained the relation between the channel capacity of a digital watermark (the amount of remaining in formation after an attack) and the distortion of the image.Study of asymmetric digital watermarkingWe proposed an idea of asymmetric digital watermarking where embedding algorithm is not the same as extracting one. That ensures the difficulty of removing a digital watermark even if the extracting method is open to the public. A concrete method was proposed and then its resistance was evaluated against some attacks.Study on efficient digital movie watermarkingWe proposed a digital watermarking method suitable for MPEG movie that requires no original image in the extraction phase. It embeds a watermark with small cost since it requires no decoding/re-encoding procedure in embedding. We evaluated its robustness against practical attacks, such as re-encoding attacks and signal processing attacks. Using patchwork technique with watermarking, it became robust against geometric attacks, such as rotation.Study on a secure content-distribution protocol and a secure traitor-tracing scheme against malicious providersWe proposed a secure information-distribution protocol that can trace a traitor who leaked the distributed information. And then, we proposed applying it to both a secure content-distribution protocol and a secure traitor tracing scheme which are secure against malicious providers.