ebook img

Imaging, Vision and Learning Based on Optimization and PDEs: IVLOPDE, Bergen, Norway, August 29 – September 2, 2016 PDF

255 Pages·2018·9.321 MB·English
Save to my drive
Quick download
Download
Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.

Preview Imaging, Vision and Learning Based on Optimization and PDEs: IVLOPDE, Bergen, Norway, August 29 – September 2, 2016

Mathematics and Visualization Xue-Cheng Tai · Egil Bae Marius Lysaker Editors Imaging, Vision and Learning Based on Optimization and PDEs IVLOPDE, Bergen, Norway, August 29 – September 2, 2016 Mathematics and Visualization Serieseditors Hans-ChristianHege DavidHoffman ChristopherR.Johnson KonradPolthier MartinRumpf Moreinformationaboutthisseriesathttp://www.springer.com/series/4562 Xue-Cheng Tai • Egil Bae (cid:129) Marius Lysaker Editors Imaging, Vision and Learning Based on Optimization and PDEs IVLOPDE, Bergen, Norway, August 29 – September 2, 2016 123 Editors Xue-ChengTai EgilBae DepartmentofMathematics NorwegianDefenceResearch UniversityofBergen Establishment Bergen,Norway Kjeller,Norway MariusLysaker UniversityofSouth-EasternNorway Porsgrunn,Norway ISSN1612-3786 ISSN2197-666X (electronic) MathematicsandVisualization ISBN978-3-319-91273-8 ISBN978-3-319-91274-5 (eBook) https://doi.org/10.1007/978-3-319-91274-5 LibraryofCongressControlNumber:2018956324 MathematicsSubjectClassification(2010):35-XX,49-XX,65-XX,90-XX ©SpringerInternationalPublishingAG,partofSpringerNature2018 Thisworkissubjecttocopyright.AllrightsarereservedbythePublisher,whetherthewholeorpartof thematerialisconcerned,specificallytherightsoftranslation,reprinting,reuseofillustrations,recitation, broadcasting,reproductiononmicrofilmsorinanyotherphysicalway,andtransmissionorinformation storageandretrieval,electronicadaptation,computersoftware,orbysimilarordissimilarmethodology nowknownorhereafterdeveloped. Theuseofgeneraldescriptivenames,registerednames,trademarks,servicemarks,etc.inthispublication doesnotimply,evenintheabsenceofaspecificstatement,thatsuchnamesareexemptfromtherelevant protectivelawsandregulationsandthereforefreeforgeneraluse. Thepublisher,theauthorsandtheeditorsaresafetoassumethattheadviceandinformationinthisbook arebelievedtobetrueandaccurateatthedateofpublication.Neitherthepublishernortheauthorsor theeditorsgiveawarranty,expressorimplied,withrespecttothematerialcontainedhereinorforany errorsoromissionsthatmayhavebeenmade.Thepublisherremainsneutralwithregardtojurisdictional claimsinpublishedmapsandinstitutionalaffiliations. ThisSpringerimprintispublishedbytheregisteredcompanySpringerNatureSwitzerlandAG Theregisteredcompanyaddressis:Gewerbestrasse11,6330Cham,Switzerland Preface It has become an established paradigm to formulate problems within image processingandcomputervisionaspartialdifferentialequations(PDEs),variational problemsorhigh-dimensionaloptimizationproblems.Thiscompact,yetexpressive framework makes it possible to incorporate a variety of desired properties of the solutions and to design algorithms based on well-founded mathematical theory. Applications range from early image formation through compressive sensing, to low-levelimageenhancement,restorationandfeaturedetection,andtohigher-level imageunderstanding,segmentationandclassification.Morerecently,thesametools have also shown great promise in more general applications of data analysis and machinelearning. InAugust2016,weorganizedaconferencetitledImaging,VisionandLearning based on Optimization and PDEs (IVLOPDE) in the city of Bergen, Norway. The conference was intended to foster collaboration and exchange of new ideas withinthesemathematicaltechniquesforthebroadrangeofapplicationsinimaging science, computer vision and machine learning. The 5-day event included invited presentations from 18 internationally leading experts within the field, and 16 contributed poster presentations. Plenty of time was also set aside for informal discussions, such as a full-day excursion to the fjords and mountains of western Norway. After the conference, the participants were invited to submit full papers based on their presentations, or based on new ideas that may have emerged during the course of the event. Each submitted article was evaluated by 2–3 expertreviewers,whoweremostlyselectedamongtheotherinvitedspeakers.The reviewersprovidedalotofusefulfeedback,whichinsomecasesledtonewinsights thattheauthorsincorporatedinrevisedversionsoftheirarticles. Thisbookconstitutestheconferencepost-proceedingsofIVLOPDE.Itcontains 11 original research articles that were selected from the submitted full papers after the peer-review process. The articles present various novel techniques and analyticalresultswithin optimization,variationalmodelsandPDEs, togetherwith experimentalresults on applications ranging from early image formation to high- level image and data analysis. To guide the reader, the articles have been divided into four topical sections: (I) image reconstruction from incomplete data, (II) v vi Preface image enhancement, restoration and registration, (III) 3D image understanding and classification, and (IV) machine learning and big data analysis. Each section featuresabalanceoftheoreticallyorientedarticlesandapplicationorientedarticles. Thebookwill benefitall researcherswithinmathematics,imagingscience ordata science who would like to become more familiar with this active field, as well as expertswhowouldliketolearnaboutthelatestdevelopments. We would like to thank all the reviewers for their valuable suggestions and careful evaluations, the Research Council of Norway (through ISP-matematikk project 239033/F20)for providingfunding for the conference, and Ruth Allewelt and Martin Peters at Springer for their excellent support and patience while we werepreparingthisbook. Bergen,Norway Xue-ChengTai Kjeller,Norway EgilBae Porsgrunn,Norway MariusLysaker April2018 Contents PartI ImageReconstructionfromIncompleteData 1 Adaptive Regularizationfor Image Reconstructionfrom SubsampledData........................................................... 3 MichaelHintermüller,AndreasLanger,CarlosN.Rautenberg, andTaoWu 2 AConvergentFixed-PointProximityAlgorithmAccelerated byFISTAforthe(cid:2) SparseRecoveryProblem.......................... 27 0 XueyingZeng,LixinShen,andYueshengXu 3 Sparse-DataBased3DSurfaceReconstruction forCartoonandMap....................................................... 47 BinWu,TalalRahman,andXue-ChengTai PartII ImageEnhancement,RestorationandRegistration 4 VariationalMethodsforGamutMapping inCinemaandTelevision.................................................. 67 SyedWaqasZamir,JavierVazquez-Corral,andMarceloBertalmío 5 FunctionalLiftingforVariationalProblemswithHigher-Order Regularization .............................................................. 101 BenediktLoewenhauserandJanLellmann 6 OntheConvexModelofSpeckleReduction............................. 121 FamingFang,YingyingFang,andTieyongZeng PartIII 3DImageUnderstandingandClassification 7 Multi-DimensionalRegularExpressionsforObjectDetection withLiDARImaging....................................................... 145 ToddC.Torgersen,V.PaúlPauca,RobertJ.Plemmons,DejanNikic, JasonWu,andRobertRand vii viii Contents 8 Relaxed Optimisation for Tensor Principal Component AnalysisandApplicationstoRecognition,Compression andRetrievalofVolumetricShapes...................................... 165 HayatoItoh,AtsushiImiya,andTomoyaSakai PartIV MachineLearningandBigDataAnalysis 9 AnIncrementalReseedingStrategyforClustering..................... 203 XavierBresson,HuiyiHu,ThomasLaurent,ArthurSzlam, andJamesvonBrecht 10 Ego-MotionClassificationforBody-WornVideos...................... 221 ZhaoyiMeng,JavierSánchez,Jean-MichelMorel, AndreaL.Bertozzi,andP.JeffreyBrantingham 11 SynchronizedRecoveryMethodforMulti-RankSymmetric TensorDecomposition...................................................... 241 HaixiaLiu,LizhangMiao,andYangWang Index............................................................................... 253 Part I Image Reconstruction from Incomplete Data

See more

The list of books you might like

Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.