Estimation of fading attack on high payload spread spectrum watermarking with variable embedding rate using genetic algorithms

Spread Spectrum (SS) watermarking is the most popular one and robust to several attacks, but the correlation receiver structure widely used for SS watermark detection is not effective in the presence of fading like attack. Moreover, watermark detection performance is greatly affected in high payload SS system under fading attack. Recently, a new model of high payload SS watermarking with minimum mean square error combining (MMSEC) decoder structure is proposed that has shown improved detection performance through multistage interference cancelation. We modify this SS watermarking for variable embedding rate so that each host signal point allows different data hiding capacity. This work further shows that the fading attack can be estimated and similar order of detection improvement can be achieved without interference cancelation or with interference cancelation at single stage. Genetic Algorithms (GAs) is used to estimate fading attack optimally for high payload SS watermark systems and this yields low bit error rate (BER) value. Simulation results show that gap in BER performance due to the actual value of fading attack and due to its estimated value is very small and this performance can be achieved with much lower computation cost. (6 pages)