Non-uniform quantizer design for image data hiding

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 (LloydMax) algorithm for vector (scalar) quantizer design. Experimental results with Barbara as the host image shows that the proposed scheme provides better tradeoffs between robustness to attacks, embedding induced distortion and embedding capacity than a simple pdf matched scheme. The proposed algorithm also shows about 3dB improvements in embedding distortion over other popular quantizer based embedding algorithms.

• Type: Conference
• Year: 2004
• Where: ICIP
• Authors: Ning Liu, Koduvayur P. Subbalakshmi
• Keywords: ICIP 2004, Image Processing, Popular Quantizer, Quantizer Based Data, Quantizer Design, Uniform Quantizer