高级检索

点云投影采样及其同步三维信息隐藏方法

Point Cloud Projection Sampling and Synchronous 3D Information Hiding

  • 摘要: 针对现有三维信息隐藏算法中顶点同步时间长和不稳定的情况,提出一种基于投影采样的同步三维点云信息隐藏方法。主要贡献是提出一种点云投影采样方法,可以消除水印嵌入坐标值后对顶点排序的不良影响,实现顶点顺序的完全同步,因此可以完全抵抗重排序攻击。该投影采样法将所有顶点垂直投影至XOY平面,通过调节投影面分辨率控制采样顶点数,然后规则存取采样点达到顶点同步。另外,本文还提出同时将秘密数据嵌入到顶点的z坐标和颜色分量中。利用提出的点云投影采样法依次将秘密数据嵌入到顶点Z轴坐标值和三个彩色分量中。由于嵌入时不改变X和Y轴坐标值,所提算法在大容量嵌入的同时保证了顶点同步的稳定性。最后,本算法提取时不需要载体模型,是一种盲提取方法。实验结果表明,这种方法容量大,不可感知性好,可以完全抵抗重排序攻击、X和Y轴的平移攻击,对X和Y轴的缩放攻击、加性噪声、剪切攻击也有一定的抵抗能力。

     

    Abstract: Aiming at the long time and unstable problems of vertex synchronization in the existing 3D information hiding algorithms, a synchronous 3D point cloud information hiding based on projection sampling is proposed. The main contribution is presenting a scheme for point cloud projection sampling, which eliminates the influence of watermark embedding on vertex ordering, and achieves complete synchronization of vertex order. So this method completely resists reordering attacks. The point cloud projection sampling method projects all vertices vertically onto the XOY plane to get the sampling points whose total number is controlled by the resolution of the projection surface. The vertex synchronization is achieved by accessing all the sampling points by rule. In addition, this paper embeds the secret data into the Z-axis coordinates and the color component values of the vertex at the same time. Due to unchanged X and Y-axis coordinates this method ensures the stability of vertex synchronization for large embedding capacity. It is a blind extraction method because the original model is not needed during the extraction. The experimental results show that this method has large capacity and good imperceptibility. It can completely resist reordering attack, X and Y-axis translation, also it has the ability to resist scaling attacks in X and Y-axis, additive noise and shear attacks.

     

/

返回文章
返回