Blind Watermarking Based on Adaptive Lattice Quantization Index Modulation

Bingwen Feng,Wei Lu,Wei Sun,Zhuoqian Liang,Juan Liu
DOI: https://doi.org/10.1007/978-3-319-31960-5_20
2015-01-01
Abstract:Lattice Quantization Index Modulation (LQIM) is an important tool in blind watermarking. Traditional compensative LQIM can only handle the global tradeoff between fidelity and robustness. To adapt the embedding strength to the local perceptual characteristics of the host signal, this paper proposes an adaptive LQIM scheme. The adaptive encoder minimizes the embedding distortion in the term of weighted-mean-squared error (wMSE) while maintaining the robustness at an acceptable level. The weight value associated with each signal element can be set according to certain perceptual measurement and is not required at the decoder. Experimental results demonstrate the superiority of the proposed scheme. Compared with the compensative LQIM, the proposed adaptive LQIM provides better fidelity without the loss of robustness.
What problem does this paper attempt to address?