High Capacity Reversible Watermarking Scheme for 2D Vector Maps

This paper presents a high capacity reversible watermarking scheme for 2D vector maps. This scheme applies an integer transform to a coordinate pair. The integer transform does not change the sum value of the coordinate pair and embeds one watermark bit. Except the first and the last one, each coordinate in an object with its previous and next one will compose two pairs respectively. Namely, each coordinate can embed one watermark bit. A threshold on difference values is predefined to distinguish the embedded pairs. As to one pair, if its absolute value of the difference is less than the threshold, the integer transform will be applied; otherwise, it is shifted. The shifted distance is larger than the embedded one so that the decoder can differentiate the two classes. To assure the distortion introduced by the transform and the shift less than the map precision distortion, the embedding condition is presented. Simulation results are provided to show that the scheme can achieve high capacity while maintaining low distortion value.