Gyrophone Recognizing Speech from Gyroscope Signals

Gyrophone Recognizing Speech from Gyroscope Signals

ID:18436004

大小:3.32 MB

页数:16页

时间:2018-09-17

Gyrophone Recognizing Speech from Gyroscope Signals_第1页
Gyrophone Recognizing Speech from Gyroscope Signals_第2页
Gyrophone Recognizing Speech from Gyroscope Signals_第3页
Gyrophone Recognizing Speech from Gyroscope Signals_第4页
Gyrophone Recognizing Speech from Gyroscope Signals_第5页
资源描述:

《Gyrophone Recognizing Speech from Gyroscope Signals》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、Gyrophone:RecognizingSpeechfromGyroscopeSignalsYanMichalevskyandDanBoneh,StanfordUniversity;GabiNakibly,NationalResearch&SimulationCenter,RafaelLtd.https://www.usenix.org/conference/usenixsecurity14/technical-sessions/presentation/michalevskyThispaperisincludedintheProceedingsofthe23rdUSENIXSecu

2、ritySymposium.August20–22,2014•SanDiego,CAISBN978-1-931971-15-7OpenaccesstotheProceedingsofthe23rdUSENIXSecuritySymposiumissponsoredbyUSENIXGyrophone:RecognizingSpeechFromGyroscopeSignalsYanMichalevskyDanBonehGabiNakiblyComputerScienceDepartmentNationalResearch&SimulationCenterStanfordUniversity

3、RafaelLtd.Abstractgyroscopesaresufficientlysensitivetomeasureacous-WeshowthattheMEMSgyroscopesfoundonmod-ticvibrations.Thisleadstothepossibilityofrecoveringernsmartphonesaresufficientlysensitivetomeasurespeechfromgyroscopereadings,namelyusingthegyro-acousticsignalsinthevicinityofthephone.There-sco

4、peasacrudemicrophone.Weshowthatthesamplingsultingsignalscontainonlyverylow-frequencyinfor-rateofthegyroscopeisupto200Hzwhichcoverssomemation(<200Hz).Neverthelessweshow,usingsignaloftheaudiblerange.Thisraisesthepossibilityofeaves-processingandmachinelearning,thatthisinformationisdroppingonspeechi

5、nthevicinityofaphonewithoutsufficienttoidentifyspeakerinformationandevenparseaccesstotherealmicrophone.speech.SinceiOSandAndroidrequirenospecialper-Asthesamplingrateofthegyroscopeislimited,onemissionstoaccessthegyro,ourresultsshowthatappscannotfullyreconstructacomprehensiblespeechfromandactiveweb

6、contentthatcannotaccessthemicro-measurementsofasinglegyroscope.Therefore,were-phonecanneverthelesseavesdroponspeechinthevicin-sorttoautomaticspeechrecognition.Weextractfea-ityofthephone.turesfromthegyroscopemeasurementsusingvarioussignalprocessingmethodsandtrainmachinelearningal-1Introductiongor

7、ithmsforrecognition.Weachieveabout50%successrateforspeakeridentificationfromasetof10speakers.Modernsmartphonesandmobiledeviceshavemanysen-Wealsoshowthatwhilelimitingourselvestoasmallvo-sorsthatenablerichuserexperience.Beingge

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。