离线灰度手写签名鉴别

离线灰度手写签名鉴别

ID:33401512

大小:4.89 MB

页数:40页

时间:2019-02-25

离线灰度手写签名鉴别_第1页
离线灰度手写签名鉴别_第2页
离线灰度手写签名鉴别_第3页
离线灰度手写签名鉴别_第4页
离线灰度手写签名鉴别_第5页
资源描述:

《离线灰度手写签名鉴别》由会员上传分享,免费在线阅读,更多相关内容在行业资料-天天文库

1、第1I页一壅圣翌垫垄堂堡圭堂垡笙奎———————————————————————————————————————————————————————————————————————一AbstractWiththepopularityofcreditcard,theattendantsecurityissuesarelncreasinglyprominent,traditionalpasswordauthenticationmethodcannotmeettheneedsofthecommunity.Usingbiologicalcharacteristics

2、toverifypersonalidentitycaneffectivelysolvethesecurityIssuesotauthentication.Signaturecharacteristicisoneofbiologicalcharacteristics.Itiswidelyusedincontracts,certificates,agreements,documentsandotherinstruments.Assomesignatureswereforged,maybecauseharmtosociety.Sotheresearchingo

3、fofflinesignaturesystem1Sveryimportant.andithaspracticalsignificanceandtheoreticalvalue.ThispaperconsidersSIFTfeaturestomostthesolebasisofthesignatureverification,whilethetramtionalmethodusethestaticfeaturesandpseudo—dynamicfeatures.Thehandwrittensi蛆atureverificationofthestatusqu

4、oathomeandabroad,aswellasthedevelopmentsareintroducedfirst.Thenputforwardaviableidentificationalgorithmbasedongray‘scalesignatureimage:inthepreprocessing,’’Quartet”isusedtodeterminesignatureimageboundary.ThenthesigmaCureimageshouldbenormalized.SubsequenttheextractionoftheSIFTfeat

5、urepointsinthegrayscaleimageareusedtothefollow—upsignatureverificationwork.MatchingSIFTf色aturepointsusingRANSACalgorithmtoeliminatethemisallocation·EachpieceofsignatureimagemayhaveanumberofSIFTfeaturesandthelocationoftheseSIFTfeatureswillbedifferent,youCannotusethematchingwayoftr

6、aditionalstaticcharacteristicsandpseudo.dynamiccharacteristics.ThisarticleextractmultiplesignaturesincommonSIFTf-eanlresasuniversalSIFTfeaturesofthesignatureanddepositethemtothesignatureSIFTfeaturelibrary.Inthefinaldecision.makingstage,usingthefeaturesintheSIFTfeaturelibrarymatch

7、withthesignaturewhichneedtobeidentify.Ifthenumberofsuccessfulmatchpointsisoveracertainamountthatthesignatureisreal.Thefinalexperimentalresultsshowthatthesystemaccuracyratereaches88.5%.,theaccuracvoftheidentificationofhigh.1evelsignaturealsoreaches80%,higheraccuracyinasimilarsigna

8、tureverificationsystem.Keywords:Offiines

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。