Image compression and watermarking using lattice vector quantization

In this paper, we propose a new algorithm which combines compression using Lattice Vector Quantization and Watermarking. It has been motivated by the fact that lossy compression tends to remove useless or non visible information while watermarking aims at inserting invisible information. We show that this combined approach enables to improve robustness to attacks like additional JPEG compression, as well as median filtering or white noise. Furthermore it is of low complexity and offers a wide range of channel capacities according to robustness.