Spread spectrum based spatial watermarking using partial embedding and reduced dynamic range detection

Abstract

In this paper, a spread spectrum based watermarking using partial embedding and reduced dynamic range detection is proposed for uncompressed video in spatial domain. Conceptually, by properly reducing the variance of host media, lower bit error rate and smaller chip-rate can be obtained. A technique of variance reduction is therefore applied by decreasing a dynamic range of the host signal. Moreover, when using the proposed technique in which some pixels having particular values are considered and modified, higher quality of the watermarked media in term of PSNR is obtained. The results show that without knowledge of original media, the embedded watermark can be recovered with a smaller chip-rate, compared to the traditional spread spectrum based methods. The improved performance can still be achieved even if various strengths of attack, i.e. addition of Gaussian distributed noise, are applied.