Topology-preserving watermarking of vector graphics Journal Article

Author(s): Huber, Stefan; Held, Martin; Meerwald, Peter; Kwitt, Roland
Article Title: Topology-preserving watermarking of vector graphics
Affiliation IST Austria
Abstract: Watermarking techniques for vector graphics dislocate vertices in order to embed imperceptible, yet detectable, statistical features into the input data. The embedding process may result in a change of the topology of the input data, e.g., by introducing self-intersections, which is undesirable or even disastrous for many applications. In this paper we present a watermarking framework for two-dimensional vector graphics that employs conventional watermarking techniques but still provides the guarantee that the topology of the input data is preserved. The geometric part of this framework computes so-called maximum perturbation regions (MPR) of vertices. We propose two efficient algorithms to compute MPRs based on Voronoi diagrams and constrained triangulations. Furthermore, we present two algorithms to conditionally correct the watermarked data in order to increase the watermark embedding capacity and still guarantee topological correctness. While we focus on the watermarking of input formed by straight-line segments, one of our approaches can also be extended to circular arcs. We conclude the paper by demonstrating and analyzing the applicability of our framework in conjunction with two well-known watermarking techniques.
Keywords: topology; distortion constraints; vector graphics; Watermarking
Journal Title: International Journal of Computational Geometry and Applications
Volume: 24
Issue 1
ISSN: 1793-6357
Publisher: World Scientific Publishing  
Date Published: 2014-03-16
Start Page: 61
End Page: 86
Copyright Statement: CC-BY 3.0
Sponsor: Work by Martin Held and Stefan Huber was supported by Austrian Science Fund (FWF): L367-N15 and P25816-N15.
DOI: 10.1142/S0218195914500034
Open access: yes (OA journal)
IST Austria Authors
  1. Stefan Huber
    11 Huber
Related IST Austria Work