Objectives:Polymerizationshrinkageofdentalcompositesremainsamajorconcerninrestorativedentistrybecauseitcanleadtomicro-crackingofthetoothanddebondingatthetooth-restorationinterface.Theaimofthisstudywastomeasurethefull-fieldpolymerizationshrinkageofdentalcompositesusingtheopticaldigitalimagecorrelation(DIC)methodandtoevaluatehowthemeasurementisinfluencedbythefactorsinexperimentsetupandimageanalysis.Methods:Fourcommercialdentalcomposites,PremiseDentine,Z100,Z250andTetricEvoCeram,weretested.Compositewasfirstplacedintoaslotmouldtoformabarspecimenwithrectangular-sectionof4mmmm,followedbythesurfacepaintingtocreateirregularspeckles.Curingwasthenappliedatoneendofthespecimenwhiletheotherpartwerecoveredagainstcuringlightforsimulatingtheclinicalcuringconditionofcompositeindentalcavity.Thepaintedsurfacewasrecordedbyacharge-coupleddevice(CCD)camerabeforeandaftercuring.Subsequently,thevolumetricshrinkageofthespecimenwascalculatedwithspecialistDICsoftwarebasedonimagecrosscorrelation.Inaddition,afewfactorsthatmayinfluencethemeasuringaccuracy,includingthesubsetwindowsize,specklesize,illuminationlightandspecimenlength,werealsoevaluated.Results:ThevolumetricshrinkageofthespecimengenerallydecreaseswithincreasingdistancefromtheirradiatedsurfacewithaconspicuousexceptionbeingthecompositePremiseDentineasitsmaximumshrinkageoccurredatasubsurfacedistanceofabout1mminsteadoftheirradiatedsurface.Z100hadthegreatestmaximumshrinkagestrain,followedbyZ250,TetricEvoCeramandthenPremiseDentine.Largersubsetwindowsizemadetheshrinkagestraincontoursmoother.Butthecostwasthatsomedetailsintheheterogeneityofthematerialwerelost.Verysmallsubsetwindowsizeresultedinalotofnoiseinthedata,makingitdifficulttodiscernthegeneralpatterninthestraindistribution.Speckle