Abstract. In audio watermarking, the robustness to desynchronization attacks such as TSM (Time-Scale Modification) operations, is still an open issue. In this paper, both mathematical proof and experimental testing show that the histogram shape (represente
Time-ScaleInvariantAudioWatermarking
BasedontheStatisticalFeatures
inTimeDomain
ShijunXiang1,2,JiwuHuang1,2,RuiYang1,2
SchoolofInformationScienceandTechnology,
SunYat-senUniversity,Guangzhou510275,China
GuangdongKeyLaboratoryofInformationSecurityTechnology,
Guangzhou510275,China
xiangshijun@gmail.com,isshjw@mail.sysu.edu.cn12
Abstract.Inaudiowatermarking,therobustnesstodesynchronizationattackssuchasTSM(Time-ScaleModi cation)operations,isstillanopenissue.Inthispaper,bothmathematicalproofandexperimentaltestingshowthatthehistogramshape(representedastherelativerelationinthenumberofsamplesamongthreedi erenthistogrambins)andtheaudiomeanaretworobustfeaturestotheTSMattacks.Accordingly,amulti-bitrobustaudiowatermarkingalgorithmbasedonthetwostatisticalfeaturesisproposedbymodifyingthehistogram.Theaudiohistogramwithequal-sizedbinsisextractedfromaselectedamplituderangereferredtotheaudiomean,andthentherelativerelationsinthenumberofsamplesamonggroupsofthreeneighboringbinsaredesignedtocarrythewatermarkbyreassigningthenumberofsamplesinthebins.Thewatermarkedaudiosignalisperceptiblysimilartotheoriginalone.SimulationresultsdemonstratedthatthehiddenmessageisveryrobusttotheTSM,cropping,andavarietyofotherdistortionsinStirmarkBenchmarkforAudio.
1Introduction
Audiowatermarking[1][2]playsanimportantroleinownershipprotection.AccordingtoIFPI(InternationalFederationofthePhonographicIndustry)[3],STEP2000[4]andSDMI(SecureDigitalMusicInitiative)[5],audiowatermarkingshouldberobusttotemporalscalingof±10%andbeabletoresistmostcommonsignalprocessingmanipulationsandattacks,suchasrandomcropping,MP3compression,resamplingandetc.
Amongthevariousproblemstobesolvedinaudiowatermarking,therobustnessagainstdesynchronizationdistortionssuchasTSMandrandomcropping,isthemostchallengingoneforpreviouswatermarkingschemesyet.Desynchronizationattacksthatcausedisplacementbetweenencoderanddecoderaredi cultforawatermarktosurvive.In[6],thesynchronizationcodewasintroducedaimingatconqueringcroppingattacks.However,thesynchronizationcodeisveryvulnerabletoTSM.Forexample,asmallamountofscaling(i.e.,
Invariant image watermarking based on statistical features in the low-frequency... Time-scale Invariant A... 暂无评价 16页 免费2014 Baidu 使用百度前...
Time-scale Invariant Aud... 暂无评价 16页 ...Invariant Domain for Image Watermarking Chaw-Seng ...is independent of image contents and its features...
A Robust Audio Watermark... 暂无评价 13页 ... A new affine invariant i... 暂无评价 10页 ... A blind watermarking scheme based on structured ...
Time-scale Invariant Aud... 暂无评价 16页 免费 IMAGE DATA COMPRESSIO......Invariant image watermarking based on statistical features in the low-frequency...
Geometric Distortions Invariant Blind Second Generation Watermarking Technique Based on Tch Zhang L, Qian GB, Xiao WW. Geometric distortions invariant blind ...
based on singular value decomposition and tiny Statistical characteristic-based ...invariant wavelet domain A robust digital audio watermarking scheme using ...
value R0 is extracted on the invariant subset. ... An algorithm for audio watermarking based on ...In the mean time, the watermarking scheme is ...
A scale interactive model based ?lter is used ...the statistical correlation between the watermark ...invariant digital image watermarking, IWDW 2002 ...
pseudo-invariant features retrieval and more speci?...As opposed to watermarking, the CBCI approach ... D. The probability of the statistical based ...
Robust Image Watermarking Using Local Invariant Features and Independent Component Analysis_专业资料。维普资讯 http://www.cqvip.com VI1 N6绷o。1o.fNaua ...
热门文档