The Multimodal Brain Tumor Image Segmentation Benchmark (BRATS)
Bjoern MenzeAsclepios Project, Inria, Sophia-Antipolis, FranceAndrás JakabETH, Computer Vision Laboratory, Zürich, SwitzerlandStefan BauerInstitute for Surgical Technology and Biomechanics, University of Bern, SwitzerlandJayashree Kalpathy–CramerDepartment of Radiology, Harvard Medical School, Boston, MA, USAKeyvan FarahaniCancer Imaging Program, National Institutes of Health, Bethesda, MD, USAJustin KirbyCancer Imaging Program, National Institutes of Health, Bethesda, MD, USAYuliya BurrenSupport Center for Advanced Neuroimaging (SCAN), Bern University Hospital, SwitzerlandNicole PorzSupport Center for Advanced Neuroimaging (SCAN), Bern University Hospital, SwitzerlandJohannes SlotboomSupport Center for Advanced Neuroimaging (SCAN), Bern University Hospital, SwitzerlandRoland WiestSupport Center for Advanced Neuroimaging (SCAN), Bern University Hospital, SwitzerlandLevente LáncziUniversity of Debrecen, Debrecen, HungaryElizabeth R. GerstnerDepartment of Neuro-oncology, Massachusetts General Hosptial, Harvard Medical School, Boston, MA, USAMarc‐André WeberDiagnostic and Interventional Radiology, University Hospital, Heidelberg, GermanyTal ArbelCentre for Intelligent Machines, McGill University, CanadaBrian AvantsDepartment of Radiology, University of Pennsylvania, Philadelphia, PA, USANicholas AyacheAsclepios Project, Inria, Sophia-Antipolis, FrancePatricia BuendiaD. Louis CollinsMcConnell Brain Imaging Centre, McGill University, CanadaNicolas CordierAsclepios Project, Inria, Sophia-Antipolis, FranceJason J. CorsoSUNY, Computer Science and Engineering, Buffalo, NY, USAAntonio CriminisiMicrosoft Research, Cambridge, UKTilak DasCambridge University Hospitals, Cambridge, UKHervé DelingetteAsclepios Project, Inria, Sophia-Antipolis, FranceÇağatay DemiralpComputer Science Department, Stanford University, Stanford, CA, USAChristopher R. DurstDepartment of Radiology and Medical Imaging, University of Virginia, Charlottesville, VA, USAMichel DojatINRIA Rhône-Alpes, Grenoble, FranceSenan DoyleINRIA Rhône-Alpes, Grenoble, FranceJoana FestaDepartment of Electronics, University Minho, PortugalFlorence ForbesINRIA Rhône-Alpes, Grenoble, FranceEzequiel GeremiaAsclepios Project, Inria, Sophia-Antipolis, FranceBen GlockerBioMedIA Group, Imperial College, London, UKPolina GollandComputer Science and Artificial Intelligence Laboratory, Massachusetts Institute of Technology, Cambridge, MA, USAXiaotao GuoDepartment of Radiology, Columbia University, New York, NY, USAAndaç HamamcıFaculty of Engineering and Natural Sciences, Sabanci University, Istanbul, TurkeyKhan M. IftekharuddinDepartment of Electrical and Computer Engineering, Old Dominion University, Norfolk, VA, USARaj JenaCambridge University Hospitals, Cambridge, UKNigel JohnDepartment of Electrical and Computer Engineering, University of Miami, Coral Gables, FL, USAEnder KonukoğluAthinoula A. Martinos Center for Biomedical Imaging, Massachusetts General Hospital and Harvard Medical School, Boston, MA, USADanial LashkariComputer Science and Artificial Intelligence Laboratory, Massachusetts Institute of Technology, Cambridge, MA, USAJosé MarizICVS/3B's—PT Government Associate Laboratory, Braga/Guimaraes, PortugalRaphael MeierInstitute for Surgical Technology and Biomechanics, University of Bern, SwitzerlandSérgio PereiraDepartment of Electronics, University Minho, PortugalDoina PrecupSchool of Computer Science, McGill University, CanadaStephen J. PriceCambridge University Hospitals, Cambridge, UKTammy Riklin RavivComputer Science and Artificial Intelligence Laboratory, Massachusetts Institute of Technology, Cambridge, MA, USASyed M. S. RezaDepartment of Electrical and Computer Engineering, Old Dominion University, Norfolk, VA, USAMichael J. RyanDuygu SarıkayaSUNY, Computer Science and Engineering, Buffalo, NY, USALawrence H. SchwartzDepartment of Radiology, Columbia University, New York, NY, USAHoo-Chang ShinElectrical and Computer Engineering Department, Ben-Gurion University, Beer-Sheva, IsraelJamie ShottonMicrosoft Research, Cambridge, UKCarlos A. SilvaDepartment of Electronics, University Minho, PortugalNuno SousaICVS/3B's—PT Government Associate Laboratory, Braga/Guimaraes, PortugalNagesh K. SubbannaDiagnostic and Interventional Radiology, University Hospital, Heidelberg, GermanyGábor SzékelyETH, Computer Vision Laboratory, Zürich, SwitzerlandThomas J. TaylorOwen ThomasCambridge University Hospitals, Cambridge, UKNicholas J. TustisonDepartment of Radiology and Medical Imaging, University of Virginia, Charlottesville, VA, USAGözde ÜnalFaculty of Engineering and Natural Sciences, Sabanci University, Istanbul, TurkeyFlor VasseurINRIA Rhône-Alpes, Grenoble, FranceMax WintermarkDepartment of Radiology and Medical Imaging, University of Virginia, Charlottesville, VA, USADong Hye YeElectrical and Computer Engineering, Purdue University, USALiang ZhaoSUNY, Computer Science and Engineering, Buffalo, NY, USABinsheng ZhaoDepartment of Radiology, Columbia University, New York, NY, USADarko ZikicMicrosoft Research, Cambridge, UKMarcel PrastawaGE Global Research, Niskayuna, NY, USAMauricio ReyesInstitute for Surgical Technology and Biomechanics, University of Bern, SwitzerlandKoen Van LeemputAalto University, Finland
2014en
ABI
Аннотация
In this paper we report the set-up and results of the Multimodal Brain Tumor Image Segmentation Benchmark (BRATS) organized in conjunction with the MICCAI 2012 and 2013 conferences. Twenty state-of-the-art tumor segmentation algorithms were applied to a set of 65 multi-contrast MR scans of low- and high-grade glioma patients-manually annotated by up to four raters-and to 65 comparable scans generated using tumor image simulation software. Quantitative evaluations revealed considerable disagreement between the human raters in segmenting various tumor sub-regions (Dice scores in the range 74%-85%), illustrating the difficulty of this task. We found that different algorithms worked best for different sub-regions (reaching performance comparable to human inter-rater variability), but that no single algorithm ranked in the top for all sub-regions simultaneously. Fusing several good algorithms using a hierarchical majority vote yielded segmentations that consistently ranked above all individual algorithms, indicating remaining opportunities for further methodological improvements. The BRATS image data and manual annotations continue to be publicly available through an online evaluation system as an ongoing benchmarking resource.
Перевод пока недоступен
Идентификаторы
Цитирования и источники
Цитирований: 6Использованных источников: 0