Robust Blind Digital 3D Model Watermarking Algorithm Using Mean Curvature

In this study, a robust and blind 3D model watermarking algorithm is introduced based on studying the geometrical properties of the 3D model. This work shows that the mean curvature (MC) has an important feature that can be used to classify the surface points. During embedding, the vertices of the surface are classified depending on the measured MC to select suitable domain for embedding and improve the imperceptibility and robustness of the watermarking algorithm. During extraction, the watermark is extracted without the need of the original 3D model. The proposed watermarking technique solves the conflict between robustness and imperceptibility by introducing watermarking algorithm that is resistant against different types of attack without affecting the quality of the watermarked model. The performance evaluation shows high imperceptibility and tolerance against many types of attack, such as noise addition, smoothing, cropping, translation, and rotation. The proposed algorithm has been designed, implemented, and tested successfully.