一种基于分块DCT变换的三维网格模型

 半脆弱水印可区分恶意攻击和正常数据处理，具有更广泛的适用性，但目前已提出的算法大多只能容忍少数类型的网格正常处理。文中提出了一种基于内容认证的半脆弱水印算法，网格分割后水印隐藏在各分块的DCT变换域系数中，网格被非法篡改时，通过匹配提取出的水印序列与原始水印序列可定位出篡改位置。实验结果表明，算法嵌入的半脆弱水印可容忍多种类型的网格正常数据处理，如网格RST相似变换和低强度的顶点坐标值量化处理，同时对恶意攻击表现敏感，可较准确的定位出篡改位置，并能以直观的可视化形式做出标记。

A Semi-fragile Watermarking Scheme for 3D Mesh Models Based on Partitioned DCT

Semi-fragile watermark has wider applicability, as it can distinguish malicious tampering and normal mesh processing. Most present algorithms can only tolerate a few types of normal mesh processing. A novel semi-fragile watermarking algorithm based on mesh content is proposed. In this algorithm, mesh is partitioned into some sub-meshes, and semi-fragile watermarks are hidden in DCT coefficients of each sub-mesh. When mesh is tampered, the position occurred can be located by matching the extracted watermark sequence and original watermark sequence. Experimental results show that the proposed algorithm can tolerate various mesh normal processing, such as mesh RST similar transforming, and quantization of vertex coordinates with low intensity, while is sensitive to malicious tampering. The proposed algorithm can also locate the tamering positions with relatively high precision.