A Highly Adaptable Capacity and Invisibility 3D Watermarking Based on Four-Points Sets

In this paper, we propose a fine tuneable capacity and invisibility novel blind watermarking method for 3D models with arbitrary topology in spatial domain. Our watermarking approach is based on a novel four-points sets construction and selection scheme to embed watermark in the vertices of the 3D models. Experimental results show that the capacity and invisibility of the watermarked models can be monotonically adjusted. To the best of our knowledge, our novel method can provide much better watermark invisibility than other state-of-the-art watermarking method, while maintain large capacity and strong robustness against various typical attacks including cropping, RST transformations and local attacks. The low computational complexity of both the embedding and the extraction processes makes it suitable for copyright protection of 3D models.