摘要： 提出一种基于SURF快速鲁棒特征点的视频分割方法，结合独立分量分析，并通过奇异值分解变换进行视频水印的嵌入，使水印在具有较好的抗时间不同步攻击和抗共谋攻击的前提下，提高对抗几何攻击的鲁棒性。实验结果表明，该算法对几何攻击、丢帧和持续时间不变减少帧数等攻击均具有较好的鲁棒性。

关键词: 视频水印, 镜头分割, SURF算法, 独立分量分析, 奇异值分解

Abstract: A video segmentation method based on Speed Up Robust Features(SURF) is proposed, which is combined with Independent Component Analysis(ICA), and through the Singular Value Decomposition(SVD) to embed the video watermark. The method has the good robust against collusion attack and temporal desynchronization, meanwhile increases the robust against geometric attacks. Experimental results show the method is robust enough against several kinds of watermark attacks, such as geometric attacks, frames discarding and decreasing.

Key words: video watermark, shot segmentation, SURF algorithm, Independent Component Analysis(ICA), Singular Value Decomposition (SVD)

中图分类号: 

• TP393