A Blind Video Watermarking Algorithm Resisting to Rotation Attack

Abstract

Rotation attack can destroy the synchronization of the watermark detector and the watermarking, even a slight rotating may invalidate the detection of watermarks. Resisting to the rotation attack, a blind video watermarking algorithm is proposed on the basis of combinding the Zernike moments and singular value decomposition (SVD). First choose the embedding areas on the luminance component. Then make the partitioning while doing the discrete wavelets transform (DWT) and make the singular value decomposition to the low-frequency parts. And last embed watermarks through modifying the maximum singular value. In detection, the possible rotation attacks to the video sequence would be corrected through Zernike moments. The experimental results show that the proposed algorithm is robust against conventional attacks and arbitrary angle of rotation attacks.