An Improved Detector for Spread-Spectrum based Watermarking using Independent Component Analysis: Performance Analysis

This paper presents a novel blind watermark detection/decoding scheme for spread spectrum (SS) based watermarking exploiting the fact that the embedded watermark and the host signal are mutually independent. The proposed detector assumes that the host signal and the watermark obey non-Gaussian distributions. The proposed scheme employs the theory of blind source separation (BSS) using independent component analysis (ICA) to cancel the host-signal interference at the detector/decoder. The paper presents analytical results showing that the proposed detector performs significantly better than the existing blind detectors used for SS-based watermarking. The paper also shows that the detection performance approaches to that of an informed detector as interference due to the host signal in the estimated watermark using BSS approaches to zero.