On optimization of spread spectrum watermarking with variable embedding rate
Santi P. Maity, Seba Maity, Jaya Sil · 2010
Several spread spectrum (SS) watermarking methods have been developed to optimize imperceptibility and robustness but majority of them suffer from capacity problem due to inherent limitation of spectrum spreading. Recently, a new model of optimized SS watermark system is proposed where the number of cover signal points, payload and watermark signal-to-interference ratio (WSIR) are optimized. The objective is to meet bit error rate (BER) and peak-to-average distortion (PAD) on individual watermarked signal point are to be retained within a predefined acceptable values under the constraint of a given global embedding distortion and cover size. We further modify this SS watermarking for variable embedding rate so that each host signal point allows different data hiding capacity leading to an overall improvement in payload. Genetic algorithms (GAs) is used to optimize this adaptive SS watermarking system so that better acceptable values for BER, PAD and per sample data hiding capacity can be achieved simultaneously.