Similarity-transformation invariant reversible watermarking for 3D models

Abstract

In this paper, we present a reversible watermarking method for 3D models with arbitrary topology based on histogram shifting. We use similarity- transformation invariants to generate histogram such that the proposed method is invariant to similarity transformations including translation, rotation and uniform scaling. By adjusting the range of a histogram adaptively, our proposed method provides a tunable capacity and invisibility for watermarking. The computational complexity of our proposed method is also very low. Experimental results demonstrate the high effectiveness of our proposed algorithm. © 2011 Springer Science+Business Media B.V.