Learning Hierarchical Features for Scene Labeling.pdf

Learning Hierarchical Features for Scene Labeling.pdf

ID:34364471

大小:1.37 MB

页数:15页

时间:2019-03-05

Learning Hierarchical Features for Scene Labeling.pdf_第1页
Learning Hierarchical Features for Scene Labeling.pdf_第2页
Learning Hierarchical Features for Scene Labeling.pdf_第3页
Learning Hierarchical Features for Scene Labeling.pdf_第4页
Learning Hierarchical Features for Scene Labeling.pdf_第5页
资源描述:

《Learning Hierarchical Features for Scene Labeling.pdf》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、1LearningHierarchicalFeaturesforSceneLabelingClementFarabet,CamilleCouprie,LaurentNajman,YannLeCun´AbstractScenelabelingconsistsinlabelingeachpixelinanimagewiththecategoryoftheobjectitbelongsto.Weproposeamethodthatusesamultiscaleconvolutionalnetworktrainedfromrawpixelstoextractdensefeaturevector

2、sthatencoderegionsofmultiplesizescenteredoneachpixel.Themethodalleviatestheneedforengineeredfeatures,andproducesapowerfulrepresentationthatcapturestexture,shapeandcontextualinformation.Wereportresultsusingmultiplepost-processingmethodstoproducethefinallabeling.Amongthose,weproposeatechniquetoauto

3、maticallyretrieve,fromapoolofsegmentationcomponents,anoptimalsetofcomponentsthatbestexplainthescene;thesecomponentsarearbitrary,e.g.theycanbetakenfromasegmentationtree,orfromanyfamilyofover-segmentations.ThesystemyieldsrecordaccuraciesontheSiftFlowDataset(33classes)andtheBarcelonaDataset(170clas

4、ses)andnear-recordaccuracyonStanfordBackgroundDataset(8classes),whilebeinganorderofmagnitudefasterthancompetingapproaches,producinga320×240imagelabelinginlessthanasecond,includingfeatureextraction.IndexTermsConvolutionalnetworks,deeplearning,imagesegmentation,imageclassification,sceneparsing.✦1IN

5、TRODUCTIONthepresenceofahumanfacegenerallyindicatesthepresenceofahumanbodynearby),butmayalsodependMAGEUNDERSTANDINGisataskofprimaryimpor-onlong-rangeinformation.Forexample,identifyingaItanceforawiderangeofpracticalapplications.Onegreypixelasbelongingtoaroad,asidewalk,agraycar,importantsteptoward

6、sunderstandinganimageistoaconcretebuilding,oracloudyskyrequiresawidecon-performafull-scenelabelingalsoknownasasceneparsing,textualwindowthatshowsenoughofthesurroundingswhichconsistsinlabelingeverypixelintheimagetomakeaninformeddecision.Toaddressthisproblem,withthecategoryoftheobjectitbelongsto.A

7、fteraweproposetouseamulti-scaleconvolutionalnetwork,perfectsceneparsing,everyregionandeveryobjectiswhichcantakeintoaccountlargeinputwindows,whiledelineatedandtagged.Onechallengeofsceneparsingkeepingthenumberoffreeparameterst

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。