Non-uniform quantizer design for image data hiding
Ning Liu, Koduvayur P. Subbalakshmi · 2005
Most quantizer based data hiding schemes use uniform quantizer, which is not optimal if the host signal is not uniformly distributed. In this paper, we design a quantizer that is not only pdf-matched but also more suited to embedding than the Linde-Buzo-Gray (Lloyd-Max) algorithm for vector (scalar) quantizer design. Experimental results shows that the proposed scheme provides better trade-offs between robustness to attacks, embedding induced distortion and embedding capacity than a simple pdf-matched scheme. The proposed algorithm also shows about 3 dB improvements in embedding distortion over other popular quantizer based embedding algorithms.