史国川　史斌浩　鲁磊纪　王　欢

(1.陆军军官学院　合肥　230031)(2.安徽医科大学　合肥　230008)



一种基于门限方案的鲁棒性视频水印算法*

史国川1史斌浩2鲁磊纪1王欢1

(1.陆军军官学院合肥230031)(2.安徽医科大学合肥230008)

摘要依据Shamir的(T,N)门限方案,提出一种鲁棒性视频水印算法。在生成水印时,采用(N,N)门限方案,将图像水印信息压缩为原信息的1/n后进行水印的嵌入,可大幅度减少水印的嵌入量。在嵌入水印时,利用DCT变换后中频系数的特性,将水印嵌入到I帧中,提高水印的鲁棒性。实验结果表明,新算法能够抵抗椒盐噪声、高斯噪声、重编码压缩等多种攻击,鲁棒性较好。

关键词门限方案; 视频水印; DCT; 鲁棒性

A Robust Video Watermarking Algorithm Based on Threshold Scheme

SHI Guochuan1SHI Binhao2LU Leiji1WANG Huan1

(1. Army Officer Academy, Hefei230031)(2. Anhui Medical University, Hefei230008)

AbstractAccording to Shamir (T,N) threshold scheme, a robust video watermarking algorithm is proposed. In the generated watermarking, using the (N,N) threshold scheme, the image watermarking information compression to the original information after the 1/Nof watermark embedding can greatly reduce the magnitude of the embedded watermarking. When embedding watermarking, characteristics of intermediate frequency coefficient after using DCT transform, embedding the watermarking into the Ⅰ frame improve the robustness of the watermarking. The experimental results show that the new algorithm can resist salt and pepper noise, Gaussian noise, re-coding compression and other many kinds of attacks, showing better robustness.

Key Wordsthreshold schemes, video watermarking, discrete cosine transformation, robustness

Class NumberTP391

1引言

目前,视频水印嵌入算法主要有基于原始视频和基于压缩视频的两种算法。基于原始视频的算法就是对未经编码的视频流数据进行处理,在原始视频中嵌入水印。基于压缩视频的算法是结合H.264等压缩标准,达到嵌入水印的目的,是目前研究的热点[1~7]。

在视频水印中追求两方面的性能:鲁棒性和不可感知性,而水印的嵌入量直接影响到水印的鲁棒性和不可感知性。一般来说,水印嵌入量越大,水印的鲁棒性越差,牺牲水印嵌入量是保证水印鲁棒性的一个有效方法。为了在视频质量和水印抗攻击性取得一个均衡,因此“先压缩再嵌入”成为数字水印嵌入的有效手法。联合的信息嵌入和有损压缩的理论分析认为[8],将压缩信息嵌入载体,在嵌入率、鲁棒性和失真上,可以实现最佳的折中办法。……