A Poisson Hidden Markov Model for Multiview Video Traffic

A Poisson Hidden Markov Model for Multiview Video Traffic

ID:40706155

大小:1.60 MB

页数:11页

时间:2019-08-06

A Poisson Hidden Markov Model for Multiview Video Traffic_第1页
A Poisson Hidden Markov Model for Multiview Video Traffic_第2页
A Poisson Hidden Markov Model for Multiview Video Traffic_第3页
A Poisson Hidden Markov Model for Multiview Video Traffic_第4页
A Poisson Hidden Markov Model for Multiview Video Traffic_第5页
资源描述:

《A Poisson Hidden Markov Model for Multiview Video Traffic》由会员上传分享,免费在线阅读,更多相关内容在学术论文-天天文库

1、1APoissonHiddenMarkovModelforMultiviewVideoTrafficLorenzoRossi,JacobChakareski,PascalFrossard,andStefaniaColonneseAbstract—Multiviewvideohasrecentlyemergedasameansresourceallocation,bufferdimensioning,andperformancetoimproveuserexperienceinnovelmultimediaservices.Weevaluation[8].proposeanews

2、tochasticmodeltocharacterizethetrafficThereishoweveralackoftrafficmodelsformultiviewgeneratedbyaMultiviewVideoCoding(MVC)variablebitvideocommunicationservices.Weproposehereanewtrafficratesource.Tothisaim,weresorttoaPoissonHiddenMarkovModel(P-HMM),inwhichthefirst(hidden)layerrepresentsthemodelfo

3、rMVCcontentinordertocharacterizetheframesizeevolutionofthevideoactivityandthesecondlayerrepresentsthesequenceobservedattheoutputofanMVCvariablebitrateframesizesofthemultipleencodedviews.Weproposeamethod(VBR)source.Specifically,buildingonourpreliminaryworkforestimatingthemodelparametersinlong

4、MVCsequences.[9],wedesignadoublystochasticsourcemodel,namelyaWethenpresentextensivenumericalsimulationsassessingthePoissonHiddenMarkovModel(P-HMM)[10],inwhichthemodel'sabilitytoproducetrafficwithrealisticcharacteristicsforageneralclassofMVCsequences.Wethenextendourfirst(hidden)layerconsistsof

5、anon-stationarychainmodelingframeworktonetworkapplicationswhereweshowthatourthevideoactivitylevelandthesecondlayerrepresentsthemodelisabletoaccuratelydescribethesenderandreceiverframesizesofthedifferentMVCencodedviews.Besides,webuffersbehaviorinMVCtransmission.Finally,wederiveaextendtheP-HM

6、Mparameterestimationalgorithmforshortmodelofuserbehaviorforinteractiveviewselection,which,inobservationsequencespresentedin[10]andadaptittolongconjunctionwithourtrafficmodel,isabletoaccuratelypredictactualnetworkloadininteractivemultiviewservices.sequencessuchasthoseencounteredinvideocommuni

7、cationservices.WeassessthemodelperformancesbyextensiveIndexTerms—Digitalvideobroadcasting,threedimensionalnumericalsimulationsonclassesofMVCsequencessharingTV,hiddenMarkovmodels,multiviewvideo.commonproperties.Weapplyourmodeltopredictthetrafficloadgenerat

当前文档最多预览五页,下载文档查看全文

此文档下载收益归作者所有

当前文档最多预览五页,下载文档查看全文
温馨提示:
1. 部分包含数学公式或PPT动画的文件,查看预览时可能会显示错乱或异常,文件下载后无此问题,请放心下载。
2. 本文档由用户上传,版权归属用户,天天文库负责整理代发布。如果您对本文档版权有争议请及时联系客服。
3. 下载前请仔细阅读文档内容,确认文档内容符合您的需求后进行下载,若出现内容与标题不符可向本站投诉处理。
4. 下载文档时可能由于网络波动等原因无法下载或下载错误,付费完成后未能成功下载的用户请联系客服处理。