Feature Based Video Watermarking Resistant to Geometric Distortions
Xiaochen Yuan, Chi‐Man Pun · 2013
This paper proposes a digital video watermarking scheme which is robust to geometric distortions, such as rotation, scaling, and cropping. The watermark is embedded / extracted based on feature extraction and local Zernike transform in / from each selected frame. The feature extraction method called Adaptive Harris Detector is proposed by adopting and revising the traditional Harris Corner Detector, and the local Zernike moments-based method is raised for watermarking use. In each selected frame, the extracted circular patches are decomposed into a collection of binary patches with Bit-Plane Decomposition method. Magnitudes of the local Zernike moments are calculated by Zernike transform and modified to embed the watermarks. Experimental results show that the proposed watermarking scheme is robust against geometric distortions and meanwhile preserves the imperceptibility of the video. Furthermore, it outperforms comparable methods when tested under common signal processing attacks.