Spread Spectrum Watermark Detection on Degraded Compressed Sensing in Matlab

This letter proposes a robust spread spectrum image watermark detection on compressed sensing (CS) measurements degraded by both multiplicative and additive noise. Watermark detection threshold is calculated first using log-likelihood ratio model. Distortion minimization problem is then formulated in terms of watermark embedding strength and the number of CS measurements under the constraint of detector reliability. A large set of simulation results show high detection probability with low false rate at low watermark power.