... types of PSF can be derived using principles of optics, if the source of theblur is known [7]. For example, out -of- focus and motion blur PSF can be parameterized with a fewparameters. Further, ... LLCImageandVideoRestorationA.MuratTekalpUniversityofRochester53.1Introduction53.2ModelingIntra-FrameObservationModel•MultispectralObserva-tionModel•MultiframeObservationModel•RegularizationModels53.3ModelParameterEstimationBlurIdentification•EstimationofRegularizationParameters•EstimationoftheNoiseVariance53.4Intra-FrameRestorationBasicRegularizedRestorationMethods•RestorationofIm-agesRecordedbyNonlinearSensors•RestorationofImagesDegradedbyRandomBlurs•AdaptiveRestorationforRing-ingReduction•BlindRestoration(Deconvolution)•Restora-tionofMultispectralImages•RestorationofSpace-VaryingBlurredImages53.5MultiframeRestorationandSuperresolutionMultiframeRestoration•Superresolution•SuperresolutionwithSpace-VaryingRestoration53.6ConclusionReferences53.1 IntroductionDigitalimagesandvideo,acquiredbystillcameras,consumercamcorders,orevenbroadcast-qualityvideocameras,areusuallydegradedbysomeamountofblurandnoise.Inaddition,mostelectroniccamerashavelimitedspatialresolutiondeterminedbythecharacteristicsofthesensorarray.Commoncausesofblurareout -of- focus,relativemotion,andatmosphericturbulence.Noisesourcesincludefilmgrain,thermal,electronic,andquantizationnoise.Further,manyimagesensorsandmediahaveknownnonlinearinput-outputcharacteristicswhichcanberepresentedaspointnonlinearities.Thegoalofimageandvideo(imagesequence)restorationistoestimateeachimage(frameorfield)asitwouldappearwithoutanydegradations,byfirstmodelingthedegradationprocess,andthenapplyinganinverseprocedure.Thisisdistinctfromimageenhancementtechniqueswhicharedesignedtomanipulateanimageinordertoproducemorepleasingresultstoanobserverwithoutmakinguseofparticulardegradationmodels.Ontheotherhand,superresolutionreferstoestimatinganimageataresolutionhigherthanthatoftheimagingsensor.Imagesequencefiltering(restorationandsuperresolution)becomesespeciallyimportantwhenstillimagesfromvideoaredesired.Thisisbecausetheblurandnoisecanbecomeratherobjectionablewhenobservinga“freeze-frame”,althoughtheymaynotbevisibletothehumaneyeattheusualframerates.Sincemanyvideosignalsencounteredinpracticeareinterlaced,weaddressthecasesofbothprogressiveandinterlacedvideo.c1999byCRCPressLLC[44] ... LLCImageandVideoRestorationA.MuratTekalpUniversityofRochester53.1Introduction53.2ModelingIntra-FrameObservationModel•MultispectralObserva-tionModel•MultiframeObservationModel•RegularizationModels53.3ModelParameterEstimationBlurIdentification•EstimationofRegularizationParameters•EstimationoftheNoiseVariance53.4Intra-FrameRestorationBasicRegularizedRestorationMethods•RestorationofIm-agesRecordedbyNonlinearSensors•RestorationofImagesDegradedbyRandomBlurs•AdaptiveRestorationforRing-ingReduction•BlindRestoration(Deconvolution)•Restora-tionofMultispectralImages•RestorationofSpace-VaryingBlurredImages53.5MultiframeRestorationandSuperresolutionMultiframeRestoration•Superresolution•SuperresolutionwithSpace-VaryingRestoration53.6ConclusionReferences53.1...