Reversible watermarking for 2D vector maps based on normalized vertices

Abstract

For 2D vector maps, obtaining rotation, uniform scaling and translation (RST) invariance property, embedding distortion control and high capacity is a technically challenging task of existing reversible watermarking schemes. Using the idea of normalized vertices, we propose a method that not only tolerates RST transformations but also provides good invisibility and high capacity. In particular, we propose transforming the vertices to a new coordinate system, calculating a normalized vertex for each transformed vertex, and embedding watermarks into the normalized vertices using an improved quantization index modulation (IQIM) based reversible watermarking technique. While the embedding into each coordinate of every normalized vertex provides RST invariance property and high capacity, the selection of the embedding parameter ensures the control of the embedding distortions. Experimental results show that the proposed scheme provides good reversibility and is robust against RST attacks. The comparisons with existing reversible watermarking schemes show that this scheme provides embedding distortion control and higher capacity.