| 摘要: |
| 目的 为了满足多用户共享三维模型版权时的单用户版权独立认证等需求，结合CDMA技术，提出一种三维点云模型高鲁棒性多重盲水印算法。方法 为不同用户分配不同的Walsh码，并利用Walsh码，对各自的二值水印图像进行编码，得到多路混合的水印；对三维点云模型进行仿射不变性处理，并将模型的顶点坐标转换为球面坐标，角度值按照升序排序，按顺序选择顶点到重心的距离组成的二维矩阵，作为水印的嵌入对象。对二维矩阵进行二级小波变换，将多路混合的水印嵌入到对角线方向高频部分，经过小波逆变换得到含多重水印的三维点云模型。结果 该算法对噪音、仿射、重排序等攻击具有很强的鲁棒性。能够嵌入多重水印，且多重水印之间没有发生相互碰撞。结论 文中算法能够满足多用户共享三维模型版权时单用户版权独立认证和版权保护的需求。 |
| 关键词: 三维点云模型 数字水印 小波变换 CDMA |
| DOI： |
| 分类号:TS853+.6 |
| 基金项目:上海市委科研创新重点项目（13ZZ111）；上海高校青年教师培养资助计划（slg14039）；上海理工大学科技发展项目（16KJFZ017） |
|
| A High Robustness Multiple Blind Watermarking Algorithm Research on 3D Point Cloud Model |
|
ZHANG Xiu-ya, SUN Liu-jie, WANG Wen-ju, QIN Yang, SHANG Jing-jing
|
|
University of Shanghai for Science and Technology, Shanghai 200093, China
|
| Abstract: |
| To meet the demand of single-user independent authentication when multi-users share copyright of 3D model, the work proposes a 3D point cloud model with a high robustness multiple blind watermarking algorithm combined with CDMA technology. Different Walsh codes were assigned to different users. Walsh codes were also used to encode individual binary watermark image. Then mixed watermark signal was obtained. The 3D point cloud model was treated by affine invariant processing. And the vertex was coordinated into spherical coordinate system. The angular values were sequenced in ascending order. Then the vertex to the center of gravity from the composition was selected as a watermark embedded object. By wavelet transform, multi-channel mixing watermark was embedded in the diagonal direction of the high frequency part. Through the inverse wavelet transform, 3D point cloud model containing multiple watermark information was obtained. The algorithm had a high robustness for noise, rotation, translation, scaling, etc., attacks. It was able to embed multiple watermarks, and no collision occurred among multiple watermarks. In conclusion, the algorithm can meet the demand of single-user independent authentication and copyright protection when multi-users share copyright of 3D model. |
| Key words: 3D point cloud model digital watermark DWT CDMA |
