Binocular stereo has been extensively studied for extracting the shape of a scene. The challenge is in matching features between two images of a scene; this is the correspondence problem. Shape from shading (SfS) is another method of extracting shape. This
IntegratingStereowithShape-from-Shading
derivedOrientationInformation
TomS.F.HainesandRichardC.Wilson
DepartmentofComputerScience
UniversityofYork
Heslington,York,UK
thaines,wilson@cs.york.ac.uk
Abstract
Binocularstereohasbeenextensivelystudiedforextractingtheshapeofascene.Thechallengeisinmatchingfeaturesbetweentwoimagesofascene;thisisthecorrespondenceproblem.Shapefromshading(SfS)isanothermethodofextractingshape.Thismodelstheinteractionoflightwiththescenesurface(s)forasingleimage.Thesetwomethodsareverydifferent;stereousessurfacefeaturestodeliveradepth-map,SfSusesshading,albedoandlightinginformationtoinferthedifferentialofthedepth-map.
Inthispaperwedevelopaframeworkfortheintegrationofbothdepthandorientationinformation.Dedicatedalgorithmsareusedforinitialestimates.AGaussian-Markovrandom eldthenrepresentsthedepth-map,GaussianbeliefpropagationisusedtoapproximatetheMAPestimateofthedepth-map.Integratinginformationfrombothstereocorrespondencesandsurfacenormalsallows nesurfacedetailstobeestimated.
1Binocularstereoisalong-standingproblemincomputervision[17].Itenablesthecon-structionof3Dmodelsfromtwo2Dimages,bysolvingthecorrespondenceproblem,wherematchingfeaturesarefoundbetweentwoimagessuchthatthematchedfeaturesareatthesamelocationontheobject.Ifcameracalibrationisthenavailabledepthcanbereconstructedfromsuchmatches.Astandardpreprocessingstepinvolvesrectifyingtheimages,sothattheyrepresenttheimagestakenfromanidealisedhorizontalparallelcamerapair.Givenarecti edimagepairfeaturescanonlymatchfeaturesonthesamescan-line.Inthedensestereoproblemwhereeverypixelisafeatureadisparitymapiscreated,wherethedisparityassignedtoeachpixelistheoffsetalongthex-axistoitsmatchingpixelintheotherimage.
Densestereoalgorithmsmaybedividedintotwosteps.Firstisthecalculationofamatchingcostforeachdisparityateachlocation,representedbythe3DDisparitySpaceImage(DSI).InareaswithstrongcuesaDSIgivesaclearindicationofactualdisparity,butinrelativelyuniformareasitwillnotdistinguishthecorrectdisparityfromincorrectdisparities.Suggestedapproachesincludenormalisedcross-correlation[17]andslidingIntroduction
Integrating Shape from Shading and Shape from Stereo...contains information on how object patches with a...orientation was not known exactly enough for our ...
depth map, we present a method for integrating the high frequency information from the shape from shading and the low frequency information from stereo. Ou...
. . . Integrating theory-driven and empirically-derived approaches to ...In particular, prototypes have a high information to time and effort ratio ...
Goal Orientation and Goal Setting: Predicting Performance by Integrating Four-...associated with deeper processing of task-related information and persistence. ...
(e.g. shape from shading, texture gradient, ...shape (i.e. surface orientation, curvature and ...ultimately obtained by integrating such information....
. . . Integrating theory-driven and empirically-derived approaches to ...In particular, prototypes have a high information to time and effort ratio ...
热门文档