手机版

Robust wide baseline stereo from maximally stable extremal r(7)

发布时间:2021-06-07   来源:未知    
字号:

The wide baseline stereo problem ,i.e.the problem of establishing correspondences between apair of images taken from different view points is studied.A new set of image elements that are put into correspondence,the so called extremal regions,is introduced.Extremal regions possess highly desirable properties

J.Matasetal./ImageandVisionComputing22(2004)761–767

767

[13]P.Pritchett,A.Zisserman,Widebaselinestereomatching,in:

Proceedingsof6thInternationalConferenceonComputerVision,Bombay,India,January1998,pp.754–760.

[14]F.Schaffalitzky,A.Zisserman,Viewpointinvarianttexturematching

andwidebaselinestereo,in:EighthInternationalConferenceonComputerVision,Vancouver,Canada,2001.

[15]F.Schaffalitzky,A.Zisserman,Multi-viewmatchingforunordered

imagesets,or‘HowdoIorganizemyholidaysnaps?’,in:ProceedingsofECCV’02,vol.1,Springer,Berlin,2002,pp.414–431.

[16]C.Schmid,R.Mohr,Localgrayvalueinvariantsforimageretrieval,

PAMI19(5)(1997)530–535.

[17]R.Sedgewick,Algorithms,seconded.,Addison-Wesley,Reading,

MA,1988.

[18]D.Tell,S.Carlsson,Widebaselinepointmatchingusingaf ne

invariantscomputedfromintensitypro les,in:ECCV’00,2000.

[19]P.Torr,A.Zisserman,Robustparameterizationandcomputationof

thetrifocaltensor,in:BMVC’96,pageMotionandActiveVision,1996.

[20]T.Tuytelaars,L.V.Gool,Content-basedimageretrievalbasedon

localaf nelyinvariantregions,in:ProceedingsofThirdInter-nationalConferenceonVisualInformationSystems,1999,pp.493–500.

[21]T.Tuytelaars,L.V.Gool,Widebaselinestereobasedonlocal,af nely

invariantregions,in:M.Mirmehdi,B.Thomas(Eds.),ProceedingsoftheBritishMachineVisionConferenceBMVC’00,London,UK,2000,pp.412–422.

[22]L.Vincent,P.Soille,Watershedsindigitalspaces:anef cient

algorithmbasedonimmersionsimulations,IEEETransactionsonPatternAnalysisandMachineIntelligence13(6)(1991)583–598.

Robust wide baseline stereo from maximally stable extremal r(7).doc 将本文的Word文档下载到电脑,方便复制、编辑、收藏和打印
×
二维码
× 游客快捷下载通道(下载后可以自由复制和排版)
VIP包月下载
特价:29 元/月 原价:99元
低至 0.3 元/份 每月下载150
全站内容免费自由复制
VIP包月下载
特价:29 元/月 原价:99元
低至 0.3 元/份 每月下载150
全站内容免费自由复制
注:下载文档有可能出现无法下载或内容有问题,请联系客服协助您处理。
× 常见问题(客服时间:周一到周五 9:30-18:00)