基于稳健特征点的平稳小波域数字水印算法

doi:10.11959/j.issn.1000-436x.2019228

通信学报 ›› 2019, Vol. 40 ›› Issue (11): 187-198.doi: 10.11959/j.issn.1000-436x.2019228

• 学术通信 • 上一篇

基于稳健特征点的平稳小波域数字水印算法

牛盼盼^1,²,杨思宇¹,王丽¹,杨红颖¹,李丽¹,王向阳¹

¹ 辽宁师范大学计算机与信息技术学院，辽宁大连 116029
² 大连理工大学电子信息与电气工程学部，辽宁大连 116023

修回日期:2019-10-21 出版日期:2019-11-25 发布日期:2019-12-06
作者简介:牛盼盼（1983- ），女，辽宁大连人，博士，辽宁师范大学副教授，主要研究方向为网络信息安全、图像处理等。|杨思宇（1995- ），女，辽宁铁岭人，辽宁师范大学硕士生，主要研究方向为信息隐藏与数字水印。|王丽（1996- ），女，辽宁大连人，辽宁师范大学硕士生，主要研究方向为信息隐藏与数字水印。|杨红颖（1965- ），女，辽宁大连人，辽宁师范大学教授，主要研究方向为多媒体信息安全、图像处理等。|李丽（1992- ），女，辽宁大连人，辽宁师范大学硕士生，主要研究方向为图像处理与数字水印。|王向阳（1965- ），男，辽宁大连人，辽宁师范大学教授，主要研究方向为网络信息安全、智能多媒体计算等。
基金资助:
国家自然科学基金资助项目(61472171);国家自然科学基金资助项目(61701212);中国博士后科学基金资助项目(2018T110220);辽宁省教育厅科学研究经费基金资助项目项目（重点攻关）(LZ2019001);辽宁省自然科学基金资助项目(2019-ZD-0468)

Digital watermarking algorithm in SWT domain based on robust local feature

Panpan NIU^1,²,Siyu YANG¹,Li WANG¹,Hongying YANG¹,Li LI¹,Xiangyang WANG¹

¹ School of Computer and Information Technology,Liaoning Normal University,Dalian 116029,China
² Faculty of Electronic Information and Electrical Engineering,Dalian University of Technology,Dalian 116023,China

Revised:2019-10-21 Online:2019-11-25 Published:2019-12-06
Supported by:
The National Natural Science Foundation of China(61472171);The National Natural Science Foundation of China(61701212);China Postdoctoral Science Foundation Funded Project(2018T110220);Key Scientific Research Project of Liaoning Provincial Department(LZ2019001);The Natural Science Foundation of Liaoning Province(2019-ZD-0468)

摘要/Abstract

摘要：

基于目前绝大多数数字音频水印方案仅仅能够对抗简单的常规信号处理，尚无法有效抵抗破坏性较强的局部去同步攻击，提出了一种基于稳健特征点的平稳小波域数字音频水印算法。首先利用高斯滤波技术计算出平稳小波域低频子带的一阶平滑梯度响应并确定基准阈值，进而提取出分布均匀且性能稳定的音频特征点；然后结合短时能量自适应确定适合于水印嵌入的局部特征音频段；最后利用扩展抖动调制方法将水印信息嵌入局部特征音频段中。仿真实验结果证明，所提算法具有较好的不可感知性，并且对常规信号处理（MP3压缩等）和去同步攻击（音调伸缩等）均具有较好的稳健性。利用描述能力强且性能稳定的平滑梯度刻画局部数字音频性质，提出一种基于平滑梯度的平稳小波域音频特征点提取方法，有效解决音频特征点稳定性差且分布极不均匀的缺点，提高了数字音频水印对幅度伸缩、局部变调、随机剪切等去同步攻击的抵抗能力。

关键词: 数字音频水印, 去同步攻击, 特征点, 平滑梯度, 平稳小波变换

Abstract:

Aiming at the challenging work to design a robust digital audio watermarking algorithm against desynchronization attacks,a new second generation digital audio watermarking in stationary wavelet transform (SWT) domain based on robust local audio feature was proposed.First,the first-order smooth gradient response of the low-pass sub-band coefficient was calculated using Gaussian filter.Then,the short-term energy was utilized to adaptively determine local feature audio segments for embedding.Finally,the watermark information was embedded into local feature audio segments with spread transform dither modulation.The experimental results show that the proposed approach has not only good transparency,but also has strong robustness against common audio processing such as MP3 compression and good robustness against the desynchronization attacks such as pitch-scale modification et al.A SWT domain audio feature point extraction method based on smooth gradient is proposed,which effectively solves the drawbacks of poor stability and uneven distribution of audio feature points,and improves the resistance of digital audio watermarks to amplitude-scale modification,pitch-scale modification,random cropping,and jittering attacks.

Key words: digital audio watermarking, desynchronization attack, feature point, smooth gradient, stationary wavelet transform

中图分类号:

TP391

牛盼盼,杨思宇,王丽,杨红颖,李丽,王向阳. 基于稳健特征点的平稳小波域数字水印算法[J]. 通信学报, 2019, 40(11): 187-198.

Panpan NIU,Siyu YANG,Li WANG,Hongying YANG,Li LI,Xiangyang WANG. Digital watermarking algorithm in SWT domain based on robust local feature[J]. Journal on Communications, 2019, 40(11): 187-198.

图/表 10

图1

图2

图3

图4

图5

图6

表1

本文算法和文献[20]算法的稳健性能对比结果"

攻击方式	本文算法	文献[20]算法	攻击方式	本文算法	文献[20]算法
低通滤波（8 kHz）	$\frac{13}{13} (100%)$	$\frac{10}{15} (66.7 %)$	幅度伸缩（+20%）	$\frac{9}{13} (69.2 %)$	$\frac{5}{15} (33.3 %)$
低通滤波（4 kHz）	$\frac{13}{13} (100%)$	$\frac{7}{13} (53.8 %)$	幅度伸缩（+10%）	$\frac{7}{13} (53.8 %)$	$\frac{7}{15} (46.7 %)$
MP3压缩（64 kbit/s）	$\frac{10}{13} (76.9 %)$	$\frac{11}{15} (73.3 %)$	幅度伸缩（-10%）	$\frac{9}{13} (69.2 %)$	$\frac{8}{15} (53.3 %)$
MP3压缩（32 kbit/s）	$\frac{4}{13} (30.8 %)$	$\frac{8}{13} (61.5 %)$	幅度伸缩（-20%）	$\frac{12}{13} (92.3 %)$	$\frac{7}{15} (46.7 %)$
重采样（44.1-11.025-44.1 kHz）	$\frac{13}{13} (100 %)$	$\frac{10}{13} (76.9 %)$	音调变化（+5%）	$\frac{13}{13} (100 %)$	$\frac{12}{15} (80 %)$
重采样（44.1-8-44.1 kHz）	$\frac{12}{13} (92.3 %)$	$\frac{10}{15} (66.7 %)$	音调变化（-5%）	$\frac{13}{13} (100 %)$	$\frac{11}{15} (73.3 %)$
重量化（16-8-16 bit）	$\frac{13}{13} (100 %)$	$\frac{11}{15} (73.3 %)$	随机剪切（2 s）	$\frac{10}{13} (76.9 %)$	$\frac{6}{15} (40 %)$
高斯白噪声（0.01）	$\frac{13}{13} (100 %)$	$\frac{8}{13} (61.5 %)$	$抖动 (\frac{1}{5 000})$	$\frac{12}{13} (92.3 %)$	$\frac{5}{15} (33.3 %)$
添加回声（100 ms,0.5）	$\frac{9}{13} (69.2 %)$	$\frac{5}{17} (29.4 %)$	TSM（110%）	$\frac{12}{13} (92.3 %)$	$\frac{4}{15} (26.7 %)$
均衡化（baseboost）	$\frac{13}{13} (100 %)$	$\frac{5}{11} (45.5 %)$	TSM（90%）	$\frac{11}{13} (84.6 %)$	$\frac{6}{15} (40 %)$

表1

表2

本文算法和文献[19]算法的稳健性能对比结果"

攻击方式		本文算法			文献[19]算法
攻击方式	Danube	Heart	Piano	Danube	Heart	Piano
MP3压缩（64 kbit/s）	$\frac{12}{13} (92.3 %)$	$\frac{9}{12} (75 %)$	$\frac{10}{13} (76.9 %)$	$\frac{14}{16} (87.5 %)$	$\frac{14}{16} (87.5 %)$	$\frac{14}{16} (87.5 %)$
MP3压缩（96 kbit/s）	$\frac{12}{13} (92.3 %)$	$\frac{11}{12} (91.7 %)$	$\frac{13}{13} (100 %)$	$\frac{15}{16} (93.8 %)$	$\frac{15}{16} (93.8 %)$	$\frac{16}{16} (100 %)$
低通滤波（4 kHz）	$\frac{13}{13} (100 %)$	$\frac{11}{12} (91.7 %)$	$\frac{13}{13} (100 %)$	$\frac{14}{16} (87.5 %)$	$\frac{8}{13} (61.5 %)$	$\frac{14}{16} (87.5 %)$
低通滤波（8 kHz）	$\frac{13}{13} (100 %)$	$\frac{12}{12} (100 %)$	$\frac{13}{13} (100 %)$	$\frac{15}{16} (93.8 %)$	$\frac{15}{16} (93.8 %)$	$\frac{16}{16} (100 %)$
均衡化	$\frac{13}{13} (100 %)$	$\frac{10}{12} (83.3 %)$	$\frac{13}{13} (100 %)$	$\frac{10}{16} (62.5 %)$	$\frac{12}{16} (75%)$	$\frac{10}{16} (62.5 %)$
添加回声	$\frac{6}{13} (46.2 %)$	$\frac{6}{12} (50 %)$	$\frac{8}{13} (61.5 %)$	$\frac{14}{16} (87.5 %)$	$\frac{15}{16} (93.8 %)$	$\frac{15}{16} (93.8 %)$
消除噪声	$\frac{12}{13} (92.3 %)$	$\frac{10}{12} (83.3 %)$	$\frac{13}{13} (100 %)$	$\frac{15}{16} (93.8 %)$	$\frac{13}{16} (81 .%)$	$\frac{16}{16} (100 %)$
重新采样（44.1-16-44.1 kHz）	$\frac{13}{13} (100 %)$	$\frac{12}{12} (100 %)$	$\frac{13}{13} (100 %)$	$\frac{16}{16} (100 %)$	$\frac{14}{16} (87.5 %)$	$\frac{16}{16} (100 %)$
重新采样（44.1-22.05-44.1 kHz）	$\frac{13}{13} (100 %)$	$\frac{12}{12} (100 %)$	$\frac{13}{13} (100 %)$	$\frac{16}{16} (100 %)$	$\frac{13}{16} (81 .3%)$	$\frac{16}{16} (100 %)$

表2

图7

图8

参考文献 27

[1]	QASIM A F , MEZIANE F , ASPIN R . Digital watermarking:applicability for developing trust in medical imaging workflows state of the art review[J]. Computer Science Review, 2018,27: 45-60.
[2]	SU Z P , YUE F , JIANG J G ,et al. SNR-constrained heuristics for optimizing the scaling parameter of robust audio watermarking[J]. IEEE Transactions on Multimedia, 2018,20(10): 2631-2644.
[3]	KAUR A , DUTTA M K . An optimized high payload audiowatermarking algorithm based on LU-factorization[J]. Multimedia Systems, 2018,24(3): 341-353.
[4]	HUA G , HUANG J , SHI Y Q ,et al. Twenty years of digital audio watermarking-a comprehensive review[J]. Signal Processing, 2016,128: 222-242.
[5]	JIANG W Z , HUANG X H , QUAN Y J . Audio watermarking algorithm against synchronization attacks using global characteristics and adaptive frame division[J]. Signal Processing, 2019,162: 153-160.
[6]	BASSIA P , PITAS I , NIKOLAIDIS N . Robust audio watermarking in the time domain[J]. IEEE Transactions on Multimedia, 2001,3(2): 232-241.
[7]	HU H T , HSU L Y . Robust,transparent and high-capacity audio watermarking in DCT domain[J]. Signal Processing, 2015,109: 226-235.
[8]	HWANG M J , LEE J S , LEE M S ,et al. SVD-based adaptive QIM watermarking on stereo audio signals[J]. IEEE Transactions on Multimedia, 2017,20(1): 45-54.
[9]	TERCHI Y , BOUGUEZEL S . A blind audio watermarking technique based on a parametric quantization index modulation[J]. Multimedia Tools and Applications, 2018,77(19): 25681-25708.
[10]	NADEAU A , SHARMA G . An audio watermark designed for efficient and robust resynchronization after analog playback[J]. IEEE Transactions on Information Forensics ＆ Security, 2017,12(6): 1393-1405.
[11]	唐鑫, 马兆丰, 钮心忻 ,等. 基于变分贝叶斯学习的音频水印盲检测方法[J]. 通信学报, 2015,36(1): 121-128.
	TANG X , MA Z F , NIU X X ,et al. Blind audio watermarking mechanism based on variational Bayesian learning[J]. Journal on Communications, 2015,36(1): 121-128.
[12]	KHALDI K , BOUDRAA A O . Audio watermarking via EMD[J]. IEEE Transactions on Audio,Speech,and Language Processing, 2009,14(12): 2619-2622.
[13]	HU H T , CHANG J R , LIN S J . Synchronous blind audio watermarking via shape configuration of sorted LWT coefficient magnitudes[J]. Signal Processing, 2018,147: 190-202.
[14]	WANG X Y , CUI Y R , YANG H Y ,et al. A new content-based digital audio watermarking algorithm for copyright protection[C]// International Conference on Information Security(Infosecu’04). 2004: 62-69.
[15]	LI W , XUE X , LU P . Localized audio watermarking technique robust against time-scale modification[J]. IEEE Transactions on Multimedia, 2006,8(1): 60-69.
[16]	MA X H , ZHANG B , DING X Y . Self-synchronization blind audio watermarking based on feature extraction and subsampling[C]// International Symposium on Neural Networks 2007,Lecture Notes in Computer Science. 2007: 40-46.
[17]	鲍德旺, 杨红颖, 祁薇 ,等. 基于音频特征的抗去同步攻击数字水印算法[J]. 中国图象图形学报, 2009,14(12): 2619-2622.
	BAO D W , YANG H Y , QI W ,et al. A content based audio watermarking against desynchronization attacks[J]. Journal of Image and Graphics, 2009,14(12): 2619-2622.
[18]	PUN C M , JIANG J J , LEE M C . Robust and self-synchronized audio watermarking by invariant feature points[J]. Information, 2012,15(2): 815-829.
[19]	PUN C M , YUAN X C . Robust segments detector for de-synchronization resilient audio watermarking[J]. IEEE Transactions on Audio Speech and Language Processing, 2013,21(11): 2412-2424.
[20]	YUAN X C , PUN C M , CHEN C L P . Robust mel-frequency cepstral coefficients feature detection and dual-tree complex wavelet transform for digital audio watermarking[J]. Information Sciences, 2015,298: 159-179.
[21]	XIANG Y , NATGUNANATHAN I , PENG D ,et al. Spread spectrum audio watermarking using multiple orthogonal PN sequences and variable embedding strengths and polarities[J]. IEEE/ACM Transactions on Audio,Speech and Language Processing, 2018,26(3): 529-539.
[22]	彭宏 . 基于机器学习方法的数字音频水印技术研究[D]. 成都:电子科技大学, 2010.
	PENG H . Digital audio watermarking techniques based on machine learning method[D]. Chengdu:University of Electronic Science and Technology of China, 2010.
[23]	张翼鹏, 陈亮, 郝欢 . 采用量子神经网络的音频水印新算法[J]. 信号处理, 2013,29(6): 684-690.
	ZHANG Y P , CHEN L , HAO H . The new audio watermarking algorithm using quantum neural networks[J]. Signal Processing, 2013,29(6): 684-690.
[24]	JLIL M , BUTT F A , MALIK A . Short-time energy,magnitude,zero crossing rate and autocorrelation measurement for discriminating voiced and unvoiced segments of speech signals[C]// IEEE International Conference on Technological Advances. 2013: 208-212.
[25]	BITAR A W , DARAZI R , COUCHOT J.F . Blind digital watermarking in PDF documents using spread transform dither modulation[J]. Multimedia Tools and Applications, 2017,76(1): 143-161.
[26]	ELLMAUTHALER A , PAGLIARI C L . Multiscale image fusion using the undecimated wavelet transform with spectral factorization and nonorthogonal filter banks[J]. IEEE Transactions on Image Processing, 2013,22(3): 1005-1017.
[27]	DESSEIN A , CONT A . An information-geometric approach to real-time audio segmentation[J]. IEEE Signal Processing Letters, 2013,20(4): 331-334.

基于稳健特征点的平稳小波域数字水印算法

Digital watermarking algorithm in SWT domain based on robust local feature

在线阅读

PDF下载

可视化

摘要/Abstract

引用本文

使用本文

图/表 10

参考文献 27

相关文章 3

Metrics

推荐阅读 0

[1]	李岩,刘念,张斌,袁开国,杨义先. 图像镜像复制粘贴篡改检测中的FI-SURF算法[J]. 通信学报, 2015, 36(5): 54-65.
[2]	王向阳,杨红颖,牛盼盼. 性能优越的混合域数字音频盲水印算法[J]. 通信学报, 2007, 28(2): 109-114.
[3]	李伟,袁一群,李晓强,薛向阳,陆佩忠. 数字音频水印技术综述[J]. 通信学报, 2005, 26(2): 100-111.