A Blind 3D Point Cloud Watermarking Algorithm Based on Azimuth Angle Modulation

Abstract

A new blind watermarking algorithm of 3D point cloud model by using azimuth angle modulation is proposed. The concave vertices with mean curvature lesser than 0 are used as feature vertices, and the rest vertices are considered to establish an invariant space for the greater robustness. The watermarking information is embedded repeatedly by modulating azimuth angle of feature vertices lying in different regions. Sub-regional embedding and redundant embedding enhance the performance of algorithm against common attacks. Experimental results demonstrate the proposed algorithm offers the better imperceptibility and robustness.