Global characteristic-based lossless watermarking for 2D-vector maps

Due to no damage to original data, the lossless watermarking is more suitable for copyright protection of vector maps. In this paper, a lossless watermarking scheme based on the global characteristics of vector map is proposed. It begins with feature point extraction of each polyline, based on which, for the extracted feature points and non-feature points, the scheme utilizes the relation model established by BP artificial neural network and the singular value decomposition (SVD) to construct the lossless watermarking parameters, respectively. Then through XOR operation between lossless parameters and copyright image, the corresponding detection keys are obtained. This scheme perfectly depicts the global characteristics of each polyline, and has complementary robustness for simplification, compression attacks and geometry attacks, achieves the comprehensive protection purpose of vector maps. Experimental results validate the complementary robustness and effective copyright distinction between different maps of the scheme. In addition, the contradiction between the robustness and imperceptibility in traditional watermarking algorithm is effectively balanced in the scheme without modifying the original vector data, so it is suitable for copyright protection of 2D-vector maps.