Optimization approach for 3D model watermarking by linear binary programming

Watermarking algorithms provide a way of hiding or embedding some bits of information in a watermark. In the case of watermarking a 3D model, many algorithms employ a so-called indexed localization scheme. In this paper, we propose an optimization framework with two new steps for such watermarking algorithms to improve their capacity and invisibility. The first step is to find an optimal layout of invariant units to improve capacity. The second step is to rearrange the correspondence between the watermark units and the invariant units to improve invisibility. Experimental tests show that by using this framework, the capacity and invisibility of watermarking algorithms can be greatly improved.