A robust digital watermarking algorithm in undecimated discrete wavelet transform domain

In this paper, a robust digital image watermarking algorithm is proposed based on fuzzy support vector machine (FSVM) correction. The significant bitplane image is extracted from original host, and undecimated discrete wavelet transform (UDWT) is performed on the significant bitplane image. The lowpass subband is divided into small blocks, and digital watermark is embedded into original host by adaptively modulating the selected wavelet coefficients in small blocks. In order to resist geometric distortion attacks, FSVM is utilized to learn geometric distortion parameters by using low order Zernike moments. The watermark extraction is carried out after watermarked image has been synchronized without original host. Simulation also shows outstanding robustness of the proposed algorithm against most common image processing operations and geometric distortions.

Graphical abstractFigure optionsDownload full-size imageDownload as PowerPoint slideHighlights► Digital watermark is embedded in UDWT domain. ► Feature vector is constructed using low order Zernike moments of significant bitplane image. ► FSVM is utilized to learn geometric distortion parameters.