Point Cloud Model Information Hiding Algorithm Based on Multi-scale Transformation and Composite Operator

In order to improve the security and robustness of the 3D model information hiding algorithm, this paper proposes a point cloud model information hiding algorithm based on multi-scale transformation and composite operator. Firstly, rasterizing the 3D point cloud model, and use the improved 3D Harris algorithm to extract the corner points of the rasterized model. Secondly, using SURF operator to screen robust feature points as embedding regions of secret information. Finally, the feature region is subjected to the multiscale transformation, and the secret information is hid by using a quantization-based method to embed it into the low-frequency coefficient matrix. The experimental results show that the algorithm can completely avoid affine transformation attacks and can achieve a Corr value of 0.729 in the face of a composite attack with 10% simplification, 0.5% noise and 10% shear. The algorithm’s invisibility, capacity, and its robustness against multiple attacks are improved.