首页 | 本学科首页   官方微博 | 高级检索  
     检索      

软判决检测的鲁棒图像水印方案
引用本文:杨文学,桑茂栋,赵耀.软判决检测的鲁棒图像水印方案[J].铁道学报,2005,27(1):45-51.
作者姓名:杨文学  桑茂栋  赵耀
作者单位:北京交通大学,信息科学研究所,北京,100044
基金项目:国家自然科学基金项目(60172062);霍英东青年教师基金项目(81053);留学回国人员科研启动基金资助
摘    要:数字水印技术是现代版权保护的手段之一,而数字水印系统的鲁棒性是其能够起到版权保护作用的必要条件,也是当前的研究热点。本文提出了一种Turbo编码的图像水印的软判决检测算法,在不改变原有的嵌入算法的前提下,只需在接收端采用软判决检测算法就能有效地提高水印系统的鲁棒性。在本文提出的方案中,首先把原始水印信息进行Turbo编码,在图像DFT变换域中,修改两个相同频点系数的大小关系来嵌入经过编码的水印信息。在检测水印时,采取软判决检测算法提取用于Turbo译码的软输入信息,可有效地提高系统的鲁棒性。实验结果表明,软判决检测的比特错误率普遍低于硬判决检测的比特错误率,在高斯噪声情况下,甚至能低0.169。

关 键 词:数字水印  软判决检测  Turbo码  离散傅立叶变换
文章编号:1001-8360(2005)01-0045-07
修稿时间:2003年12月22

Soft-decision Detection Robust Image Watermarking Scheme
YANG Wen-xue,SANG Mao-dong,ZHAO Yao.Soft-decision Detection Robust Image Watermarking Scheme[J].Journal of the China railway Society,2005,27(1):45-51.
Authors:YANG Wen-xue  SANG Mao-dong  ZHAO Yao
Abstract:Digital watermarking technology is one of the means of copyright protection. Robustness of a digital watermarking system is a necessity to protect digital works' copyrights, and it is the hot spot of research at present. In this paper, a soft detection algorithm for Turbo-coded robust image watermarking scheme is proposed. Robustness of a watermarking system is improved without any change to watermark embedding when we apply the soft detection algorithm at the receiving end. Watermark information is encoded by Turbo-codes firstly and then is embedded into the original image in the DFT domain. For each bit of the coded information, a pair of points with the same frequency is pseudo-randomly selected and modified. During watermark detection, softinput messages for Turbo decoding are collected by adopting the soft-decision detection algorithm, which improves the robustness of the watermarking system effectively. Experimental results indicate that the Bit Error Rate (BER) of soft-decision detection femains lower than that of hard-decision detection, even 0.169 lower in gaussian noises.
Keywords:digital watermark  soft-decision  turbo-codes  Discrete Fourier ransform (DFT)  
本文献已被 CNKI 维普 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号