检索结果-内蒙古大学图书馆

Naive gabor networks for hyperspectral image classification

学校读者我要写书评

暂无评论

arXiv 2019年

作者： Liu, Chenying Li, Jun He, Lin Plaza, Antonio Li, Shutao Li, Bo Guangdong Provincial Key Laboratory of Urbanization and Geo-simulation School of Geography and Planning Sun Yat-sen University Guangzhou510275 China School of Automation Science and Engineering South China University of Technology Guangzhou510640 China College of Electrical and Information Engineering Hunan University Changsha410082 China Hyperspectral Computing Laboratory Department of Technology of Computers and Communications Escuela Politécnica University of Extremadura CáceresE-10071 Spain Beijing Key Laboratory of Digital Media School of Computer Science and Engineering State Key Laboratory of Virtual Reality Technology and Systems Beihang University Beijing100191 China

Recently, many convolutional neural network (CNN) methods have been designed for hyperspectral image (HSI) classification since CNNs are able to produce good representations of data, which greatly benefits from a huge number of parameters. However, solving such a high-dimensional optimization problem often requires a large amount of training samples in order to avoid overfitting. Additionally, it is a typical non-convex problem affected by many local minima and flat regions. To address these problems, in this paper, we introduce naive Gabor Networks or Gabor-Nets which, for the first time in the literature, design and learn CNN kernels strictly in the form of Gabor filters, aiming to reduce the number of involved parameters and constrain the solution space, and hence improve the performances of CNNs. Specifically, we develop an innovative phase-induced Gabor kernel, which is trickily designed to perform the Gabor feature learning via a linear combination of local low-frequency and high-frequency components of data controlled by the kernel phase. With the phase-induced Gabor kernel, the proposed Gabor-Nets gains the ability to automatically adapt to the local harmonic characteristics of the HSI data and thus yields more representative harmonic features. Also, this kernel can fulfill the traditional complex-valued Gabor filtering in a real-valued manner, hence making Gabor-Nets easily perform in a usual CNN thread. We evaluated our newly developed Gabor-Nets on three well-known HSIs, suggesting that our proposed Gabor-Nets can significantly improve the performance of CNNs, particularly with a small training set. Copyright © 2019, The Authors. All rights reserved.

关键词： Spectroscopy

Search for Axionlike Dark Matter Using Solid-State Nuclear Magnetic Resonance

学校读者我要写书评

暂无评论

Physical Review Letters 2021年第14期126卷 141802-141802页

作者： Deniz Aybas Janos Adam Emmy Blumenthal Alexander V. Gramolin Dorian Johnson Annalies Kleyheeg Samer Afach John W. Blanchard Gary P. Centers Antoine Garcon Martin Engler Nataniel L. Figueroa Marina Gil Sendra Arne Wickenbrock Matthew Lawson Tao Wang Teng Wu Haosu Luo Hamdi Mani Philip Mauskopf Peter W. Graham Surjeet Rajendran Derek F. Jackson Kimball Dmitry Budker Alexander O. Sushkov Department of Physics Boston University Boston Massachusetts 02215 USA Department of Electrical and Computer Engineering Boston University Boston Massachusetts 02215 USA Helmholtz-Institut GSI Helmholtzzentrum für Schwerionenforschung 55128 Mainz Germany Johannes Gutenberg-Universität Mainz 55128 Mainz Germany The Oskar Klein Centre for Cosmoparticle Physics Department of Physics Stockholm University AlbaNova 10691 Stockholm Sweden Nordita KTH Royal Institute of Technology and Stockholm University Roslagstullsbacken 23 10691 Stockholm Sweden Department of Physics Princeton University Princeton New Jersey 08544 USA State Key Laboratory of Advanced Optical Communication Systems and Networks Department of Electronics and Center for Quantum Information Technology Peking University Beijing 100871 China Shanghai Institute of Ceramics Chinese Academy of Sciences Shanghai 201800 China School of Earth and Space Exploration Arizona State University Tempe Arizona 85287 USA Stanford Institute for Theoretical Physics Stanford University Stanford California 94305 USA Department of Physics and Astronomy The Johns Hopkins University Baltimore Maryland 21218 USA Department of Physics California State University–East Bay Hayward California 94542-3084 USA Department of Physics University of California Berkeley California 94720-7300 USA Photonics Center Boston University Boston Massachusetts 02215 USA

We report the results of an experimental search for ultralight axionlike dark matter in the mass range 162–166 neV. The detection scheme of our Cosmic Axion Spin Precession Experiment is based on a precision measurement of Pb207 solid-state nuclear magnetic resonance in a polarized ferroelectric crystal. Axionlike dark matter can exert an oscillating torque on Pb207 nuclear spins via the electric dipole moment coupling gd or via the gradient coupling gaNN. We calibrate the detector and characterize the excitation spectrum and relaxation parameters of the nuclear spin ensemble with pulsed magnetic resonance measurements in a 4.4 T magnetic field. We sweep the magnetic field near this value and search for axionlike dark matter with Compton frequency within a 1 MHz band centered at 39.65 MHz. Our measurements place the upper bounds |gd|<9.5×10−4 GeV−2 and |gaNN|<2.8×10−1 GeV−1 (95% confidence level) in this frequency range. The constraint on gd corresponds to an upper bound of 1.0×10−21 e cm on the amplitude of oscillations of the neutron electric dipole moment and 4.3×10−6 on the amplitude of oscillations of CP-violating θ parameter of quantum chromodynamics. Our results demonstrate the feasibility of using solid-state nuclear magnetic resonance to search for axionlike dark matter in the neV mass range.

关键词： Particle dark matter Axions Ferroelectrics Relaxor ferroelectrics Dark matter detectors Nuclear magnetic resonance Precision measurements

Understanding adversarial attacks on deep learning based medical image analysis systems

学校读者我要写书评

暂无评论

arXiv 2019年

作者： Ma, Xingjun Niu, Yuhao Gu, Lin Wang, Yisen Zhao, Yitian Bailey, James Lu, Feng State Key Laboratory of VR Technology and Systems School of CSE Beihang University Beijing China School of Computing and Information Systems The University of Melbourne ParkvilleVIC3010 Australia Beijing Advanced Innovation Center for Big Data-Based Precision Medicine Beihang University Beijing China National Institute of Informatics Tokyo101-8430 Japan Department of Computer Science and Engineering Shanghai Jiao Tong University Shanghai China Cixi Instuitue of Biomedical Engineering Ningbo Institute of Industrial Technology Chinese Academy of Sciences Ningbo China

Deep neural networks (DNNs) have become popular for medical image analysis tasks like cancer diagnosis and lesion detection. However, a recent study demonstrates that medical deep learning systems can be compromised by carefully-engineered adversarial examples/attacks with small imperceptible perturbations. This raises safety concerns about the deployment of these systems in clinical settings. In this paper, we provide a deeper understanding of adversarial examples in the context of medical images. We find that medical DNN models can be more vulnerable to adversarial attacks compared to models for natural images, according to two different viewpoints. Surprisingly, we also find that medical adversarial attacks can be easily detected, i.e., simple detectors can achieve over 98% detection AUC against state-of-the-art attacks, due to fundamental feature differences compared to normal examples. We believe these findings may be a useful basis to approach the design of more explainable and secure medical deep learning systems. Copyright © 2019, The Authors. All rights reserved.

关键词： Deep neural networks

Common Limitations of Image Processing Metrics: A Picture Story

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Reinke, Annika Tizabi, Minu D. Sudre, Carole H. Eisenmann, Matthias Rädsch, Tim Baumgartner, Michael Acion, Laura Antonelli, Michela Arbel, Tal Bakas, Spyridon Bankhead, Peter Benis, Arriel Blaschko, Matthew Buettner, Florian Cardoso, M. Jorge Chen, Jianxu Cheplygina, Veronika Christodoulou, Evangelia Cimini, Beth A. Collins, Gary S. Engelhardt, Sandy Farahani, Keyvan Ferrer, Luciana Galdran, Adrian van Ginneken, Bram Glocker, Ben Godau, Patrick Haase, Robert Hamprecht, Fred Hashimoto, Daniel A. Heckmann-Nötzel, Doreen Hirsch, Peter Hoffman, Michael M. Huisman, Merel Isensee, Fabian Jannin, Pierre Kahn, Charles E. Kainmueller, Dagmar Kainz, Bernhard Karargyris, Alexandros Karthikesalingam, Alan Kavur, A. Emre Kenngott, Hannes Kleesiek, Jens Kleppe, Andreas Koehler, Sven Kofler, Florian Kopp-Schneider, Annette Kooi, Thijs Kozubek, Michal Kreshuk, Anna Kurc, Tahsin Landman, Bennett A. Litjens, Geert Madani, Amin Maier-Hein, Klaus Martel, Anne L. Mattson, Peter Meijering, Erik Menze, Bjoern Moher, David Moons, Karel G.M. Müller, Henning Nichyporuk, Brennan Nickel, Felix Noyan, M. Alican Petersen, Jens Polat, Gorkem Rafelski, Susanne M. Rajpoot, Nasir Reyes, Mauricio Rieke, Nicola Riegler, Michael A. Rivaz, Hassan Saez-Rodriguez, Julio Sánchez, Clara I. Schroeter, Julien Saha, Anindo Selver, M. Alper Sharan, Lalith Shetty, Shravya Smeden, Maarten V.A.N. Stieltjes, Bram Summers, Ronald M. Taha, Abdel A. Tiulpin, Aleksei Tsaftaris, Sotirios A. Calster, Ben V.A.N. Varoquaux, Gaël Wiesenfarth, Manuel Yaniv, Ziv R. Jäger, Paul Maier-Hein, Lena Division of Intelligent Medical Systems and HI Helmholtz Imaging Heidelberg Germany Faculty of Mathematics and Computer Science Heidelberg University Heidelberg Germany Division of Intelligent Medical Systems Heidelberg Germany NCT Heidelberg DKFZ University Medical Center Heidelberg Germany MRC Unit for Lifelong Health and Ageing UCL Centre for Medical Image Computing Department of Computer Science University College London London United Kingdom School of Biomedical Engineering and Imaging Science King’s College London London United Kingdom Division of Medical Image Computing Heidelberg Germany Instituto de Cálculo CONICET – Universidad de Buenos Aires Buenos Aires Argentina Centre for Medical Image Computing University College London London United Kingdom McGill University Montréal Canada Division of Computational Pathology Dept of Pathology & Laboratory Medicine Indiana University School of Medicine IU Health Information and Translational Sciences Building Indianapolis United States University of Pennsylvania Richards Medical Research Laboratories FL7 PhiladelphiaPA United States Institute of Genetics and Cancer University of Edinburgh Edinburgh United Kingdom Department of Digital Medical Technologies Holon Institute of Technology Holon Israel European Federation for Medical Informatics Le Mont-sur-Lausanne Switzerland Center for Processing Speech and Images Department of Electrical Engineering KU Leuven Kasteelpark Arenberg 10 - box 2441 Leuven3001 Belgium Frankfurt/Mainz DKFZ UCT Frankfurt-Marburg Germany Heidelberg Germany Goethe University Frankfurt Department of Medicine Germany Goethe University Frankfurt Department of Informatics Germany Frankfurt Cancer Insititute Germany Leibniz-Institut für Analytische Wissenschaften – ISAS – e.V. Dortmund Germany Department of Computer Science IT University of Copenhagen Copenhagen Denmark Imaging Platform Broad Institute of MIT and Harvard CambridgeMA United States Centre for St

While the importance of automatic image analysis is continuously increasing, recent meta-research revealed major flaws with respect to algorithm validation. Performance metrics are particularly key for meaningful, objective, and transparent performance assessment and validation of the used automatic algorithms, but relatively little attention has been given to the practical pitfalls when using specific metrics for a given image analysis task. These are typically related to (1) the disregard of inherent metric properties, such as the behaviour in the presence of class imbalance or small target structures, (2) the disregard of inherent data set properties, such as the non-independence of the test cases, and (3) the disregard of the actual biomedical domain interest that the metrics should reflect. This living dynamically document has the purpose to illustrate important limitations of performance metrics commonly applied in the field of image analysis. In this context, it focuses on biomedical image analysis problems that can be phrased as image-level classification, semantic segmentation, instance segmentation, or object detection task. The current version is based on a Delphi process on metrics conducted by an international consortium of image analysis experts from more than 60 institutions worldwide. © 2021, CC BY.

关键词： Medical imaging

Realizing topologically ordered states on a quantum processor

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Satzinger, Kevin J. Liu, Y. Smith, A. Knapp, C. Newman, M. Jones, C. Chen, Z. Quintana, C. Mi, X. Dunsworth, A. Gidney, C. Aleiner, I. Arute, F. Arya, K. Atalaya, J. Babbush, R. Bardin, J.C. Barends, R. Basso, J. Bengtsson, A. Bilmes, A. Broughton, M. Buckley, B.B. Buell, D.A. Burkett, B. Bushnell, N. Chiaro, B. Collins, R. Courtney, W. Demura, S. Derk, A.R. Eppens, D. Erickson, C. Farhi, E. Foaro, L. Fowler, A.G. Foxen, B. Giustina, M. Greene, A. Gross, J.A. Harrigan, M.P. Harrington, S.D. Hilton, J. Hong, S. Huang, T. Huggins, W.J. Ioffe, L.B. Isakov, S.V. Jeffrey, E. Jiang, Z. Kafri, D. Kechedzhi, K. Khattar, T. Kim, S. Klimov, P.V. Korotkov, A.N. Kostritsa, F. Landhuis, D. Laptev, P. Locharla, A. Lucero, E. Martin, O. McClean, J.R. McEwen, M. Miao, K.C. Mohseni, M. Montazeri, S. Mruczkiewicz, W. Mutus, J. Naaman, O. Neeley, M. Neill, C. Niu, M.Y. O'Brien, T.E. Opremcak, A. Pató, B. Petukhov, A. Rubin, N.C. Sank, D. Shvarts, V. Strain, D. Szalay, M. Villalonga, B. White, T.C. Yao, Z. Yeh, P. Yoo, J. Zalcman, A. Neven, H. Boixo, S. Megrant, A. Chen, Y. Kelly, J. Smelyanskiy, V. Kitaev, A. Knap, M. Pollmann, F. Roushan, P. Google Quantum AI Mountain ViewCA United States Department of Physics Technical University of Munich Garching85748 Germany Schellingstr. 4 München80799 Germany School of Physics and Astronomy University of Nottingham NottinghamNG7 2RD United Kingdom Centre for the Mathematics and Theoretical Physics of Quantum Non-Equilibrium Systems University of Nottingham NottinghamNG7 2RD United Kingdom Department of Physics Institute for Quantum Information and Matter California Institute of Technology PasadenaCA United States Walter Burke Institute for Theoretical Physics California Institute of Technology PasadenaCA United States Department of Electrical and Computer Engineering University of Massachusetts AmherstMA United States Laboratoire de Physique Theorique et Hautes Energies Sorbonne Université France Research Laboratory of Electronics Massachusetts Institute of Technology CambridgeMA02139 United States Department of Physics University of California Santa BarbaraCA United States Institute for Advanced Study Technical University of Munich Garching85748 Germany

The discovery of topological order has revolutionized the understanding of quantum matter in modern physics and provided the theoretical foundation for many quantum error correcting codes. Realizing topologically ordered states has proven to be extremely challenging in both condensed matter and synthetic quantum systems. Here, we prepare the ground state of the toric code Hamiltonian using an efficient quantum circuit on a superconducting quantum processor. We measure a topological entanglement entropy near the expected value of ln 2, and simulate anyon interferometry to extract the braiding statistics of the emergent excitations. Furthermore, we investigate key aspects of the surface code, including logical state injection and the decay of the non-local order parameter. Our results demonstrate the potential for quantum processors to provide key insights into topological quantum matter and quantum error correction. © 2021, CC BY.

关键词： Topology

Building a quantum engineering undergraduate program

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Asfaw, Abraham Blais, Alexandre Brown, Kenneth R. Candelaria, Jonathan Cantwell, Christopher Carr, Lincoln D. Combes, Joshua Debroy, Dripto M. Donohue, John M. Economou, Sophia E. Edwards, Emily Fox, Michael F.J. Girvin, Steven M. Ho, Alan Hurst, Hilary M. Jacob, Zubin Johnson, Blake R. Johnston-Halperin, Ezekiel Joynt, Robert Kapit, Eliot Klein-Seetharaman, Judith Laforest, Martin Lewandowski, H.J. Lynn, Theresa W. McRae, Corey Rae H. Merzbacher, Celia Michalakis, Spyridon Narang, Prineha Oliver, William D. Palsberg, Jens Pappas, David P. Raymer, Michael G. Reilly, David J. Saffman, Mark Searles, Thomas A. Shapiro, Jeffrey H. Singh, Chandralekha IBM Quantum IBM T. J. Watson Research Center Yorktown HeightsNY United States Institut Quantique Département de Physique Université de Sherbrooke SherbrookeQCJ1K 2R1 Canada Canadian Institute for Advanced Research TorontoONM5G 1M1 Canada The Duke Quantum Center Department of Physics Department of Electrical and Computer Engineering Department of Chemistry Duke University DurhamNC27708 United States SystemX Program Department of Electrical Engineering Stanford University StanfordCA94305 United States Department of Physics and Astronomy University of Southern California Los AngelesCA90089 United States The Quantum Engineering Program Department of Physics Colorado School of Mines GoldenCO80401 United States The Department of Electrical Computer and Energy Engineering University of Colorado Boulder BoulderCO80309 United States The Institute for Quantum Computing University of Waterloo WaterlooONN2L 3G1 Canada The Department of Physics Virginia Tech BlacksburgVA24061 United States The IQUIST University of Illinois Urbana-Champaign UrbanaIL61801 United States JILA National Institute of Standards and Technology The University of Colorado BoulderCO80309 United States Department of Physics University of Colorado Boulder BoulderCO80309 United States The Department of Physics Imperial College London Prince Consort Road LondonSW7 2AZ United Kingdom The Yale Quantum Institute Department of Physics Yale University New HavenCT06520 United States Google Research VeniceCA90291 United States The Department of Physics & Astronomy San José State University San JoséCA95192 United States The School of Electrical and Computer Engineering Purdue University Indiana West Lafayatte47907 United States The Department of Physics The Ohio State University ColumbusOH43210 United States The Department of Physics University of Wisconsin-Madison 1150 University Avenue MadisonWI53706 United States The Quantitative Biosciences and Engineering

The rapidly growing quantum information science and engineering (QISE) industry will require both quantum-aware and quantum-proficient engineers at the bachelor's level. We provide a roadmap for building a quantum engineering education program to satisfy this need. For quantum-aware engineers, we describe how to design a first quantum engineering course accessible to all STEM students. For the education and training of quantum-proficient engineers, we detail both a quantum engineering minor accessible to all STEM majors, and a quantum track directly integrated into individual engineering majors. We propose that such programs typically require only three or four newly developed courses that complement existing engineering and science classes available on most larger campuses. We describe a conceptual quantum information science course for implementation at any post-secondary institution, including community colleges and military schools. QISE presents extraordinary opportunities to work towards rectifying issues of inclusivity and equity that continue to be pervasive within engineering. We present a plan to do so and describe how quantum engineering education presents an excellent set of education research opportunities. Finally, we outline a hands-on training plan on quantum hardware, a key component of any quantum engineering program, with a variety of technologies including optics, atoms and ions, cryogenic and solid-state technologies, nanofabrication, and control and readout electronics. Our recommendations provide a flexible framework that can be tailored for academic institutions ranging from teaching and undergraduate-focused two- and four-year colleges to research-intensive universities. © 2021, CC BY.

关键词： engineering education

Generalized superradiant assembly for nanophotonic thermal emitters

学校读者我要写书评

暂无评论

Physical Review B 2018年第12期97卷 125406-125406页

作者： Sudaraka Mallawaarachchi Sarath D. Gunapala Mark I. Stockman Malin Premaratne Advanced Computing and Simulation Laboratory (AχL) Department of Electrical and Computer Systems Engineering Monash University Clayton Victoria 3800 Australia Jet Propulsion Laboratory California Institute of Technology Pasadena California 91109 USA Department of Physics and Astronomy Georgia State University Atlanta Georgia 30303 USA

Superradiance explains the collective enhancement of emission, observed when nanophotonic emitters are arranged within subwavelength proximity and perfect symmetry. Thermal superradiant emitter assemblies with variable photon far-field coupling rates are known to be capable of outperforming their conventional, nonsuperradiant counterparts. However, due to the inability to account for assemblies comprising emitters with various materials and dimensional configurations, existing thermal superradiant models are inadequate and incongruent. In this paper, a generalized thermal superradiant assembly for nanophotonic emitters is developed from first principles. Spectral analysis shows that not only does the proposed model outperform existing models in power delivery, but also portrays unforeseen and startling characteristics during emission. These electromagnetically induced transparency like (EIT-like) and superscattering-like characteristics are reported here for a superradiant assembly, and the effects escalate as the emitters become increasingly disparate. The fact that the EIT-like characteristics are in close agreement with a recent experimental observation involving the superradiant decay of qubits strongly bolsters the validity of the proposed model.

关键词： Nanophotonics Superradiance & subradiance

The association for the advancement of artificial intelligence 2020 workshop program

学校读者我要写书评

暂无评论

The association for the advancement of artificial intelligen...

作者： Bang, Grace Barash, Guy Bea, Ryan Cali, Jacques Castillo-Effen, Mauricio Chen, Xin Cynthia Chhaya, Niyati Dhoopar, Rohan Dumancic, Sebastijan Espinoza, Huáscar Farchi, Eitan Fioretto, Ferdinando Fuentetaja, Raquel Galley, Michel Geib, Christopher Hernández-Orallo, José Huang, Xiaowei Keren, Sarah Kim, Seokhwan Jaidka, Kokil Liu, Xiaomo Lu, Tyler Ma, Zhiqiang Mallah, Richard McDermid, John Michalowski, Martin Mirsky, Reuth Héigeartaigh, Seán Ó. Ramachandran, Deepak Segovia-Aguas, Javier Shaban-Nejad, Arash Shehory, Onn Shwartz, Vered Srivastava, Siddharth Talamadupula, Kartik Tang, Jian Zhang, Dell Zhang, Jian S&P Global. Western Digital University of Southampton United Kingdom Blue Prism Group PLC Lockheed Martin's Advanced Technology Laboratories ArlingtonVA United States University of Hong Kong Hong Kong Adobe Research Google Research Katholieke Universiteit Leuven Belgium Commissariat à ĺÉnergie Atomique France IBM Research Department of Electrical Engineering and Computer Science Syracuse University United States Universidad Carlos III de Madrid Spain Microsoft Research AI SIFT LLC Universitat Politècnica de València Spain Leverhulme Centre for the Future of Intelligence United Kingdom Department of Computer Science University of Liverpool United Kingdom National University of Singapore Singapore Harvard School of Engineering and Applied Sciences in the of Fields of Multiagent Environment Design and Goal Recognition Design Amazon Alexa AI Future of Life Institute United States University of York Assuring Autonomy International Program United Kingdom University of Minnesota School of Nursing MinneapolisMN United States University of Texas at Austin AustinTX United States Institut de Robòtica i Informàtica Industrial Spanish Council for Scientific Research Technical University of Catalonia Spain Oak Ridge National Laboratory Center for Biomedical Informatics Department of Pediatrics University of Tennessee Health Science Center United States Bar Ilan University Israel University of Washington SeattleWA United States Autonomous Agents and Intelligent Robots lab of the School of Computing Informatics Decision Systems Arizona State University United States DiDi AI Labs School of Industrial and System Engineering Georgia Institute of Technology United States University of London United Kingdom Microsoft Azure

The Association for the Advancement of Artificial Intelligence 2020 Workshop Program included twenty-three workshops covering a wide range of topics in artificial intelligence. This report contains the required reports, which were submitted by most, but not all, of the workshop chairs. Copyright © 2020, Association for the Advancement of Artificial Intelligence. All rights reserved.

关键词：

The medical segmentation decathlon

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Antonelli, Michela Reinke, Annika Bakas, Spyridon Farahani, Keyvan Kopp-Schneider, Annette Landman, Bennett A. Litjens, Geert Menze, Bjoern Ronneberger, Olaf Summers, Ronald M. van Ginneken, Bram Bilello, Michel Bilic, Patrick Christ, Patrick F. Do, Richard K.G. Gollub, Marc J. Heckers, Stephan H. Huisman, Henkjan Jarnagin, William R. McHugo, Maureen K. Napel, Sandy Golia Pernicka, Jennifer S. Rhode, Kawal Tobon-Gomez, Catalina Vorontsov, Eugene Hulsmon, Hankan Meakin, James A. Ourselin, Sebastien Wiesenfarth, Manuel Arbeláez, Pablo Bae, Byeonguk Chen, Sihong Daza, Laura Feng, Jianjiang He, Baochun Isensee, Fabian Ji, Yuanfeng Jia, Fucang Kim, Namkug Kim, Ildoo Merhof, Dorit Pai, Akshay Park, Beomhee Perslev, Mathias Rezaiifar, Ramin Rippel, Oliver Sarasua, Ignacio Shen, Wei Son, Jaemin Wachinger, Christian Wang, Liansheng Wang, Yan Xia, Yingda Xu, Daguang Xu, Zhanwei Zheng, Yefeng Simpson, Amber L. Maier-Hein, Lena Cardoso, M. Jorge School of Biomedical Engineering & Imaging Sciences King’s College London London United Kingdom Heidelberg Germany Heidelberg Germany Faculty of Mathematics and Computer Science University of Heidelberg Heidelberg Germany University of Pennsylvania PhiladelphiaPA United States BethesdaMD United States Heidelberg Germany Electrical Engineering and Computer Science Vanderbilt University NashvilleTN United States Radboud University Medical Center Radboud Institute for Health Sciences Nijmegen Netherlands Quantitative Biomedicine University of Zurich Zurich Switzerland DeepMind London United Kingdom School of Computing Department of Biomedical and Molecular Sciences Queen’s University KingstonON Canada Department of Surgery Memorial Sloan Kettering Cancer Center New YorkNY United States BethesdaMD United States Department of Informatics Technische Universität München Germany Department of Radiology Memorial Sloan Kettering Cancer Center New YorkNY United States Department of Psychiatry & Behavioral Sciences Vanderbilt University Medical Center NashvilleTN United States Department of Radiology Stanford University StanfordCA United States Department of Computer Science and Software Engineering École Polytechnique de Montréal Department of Radiology Perelman School of Medicine University of Pennsylvania PhiladelphiaPA United States Department of Pathology and Laboratory Medicine Perelman School of Medicine University of Pennsylvania PhiladelphiaPA United States Medical Faculty University of Heidelberg Heidelberg Germany VUNO Inc. Seoul Korea Republic of Universidad de los Andes Colombia Department of Automation Tsinghua University Beijing China Heidelberg Germany Department of Computer Science Xiamen University Xiamen361005 China Shenzhen Institute of Advanced Technology Chinese Academy of Sciences Shenzhen China Kakao Brain Korea Republic of Institute of Imaging & Computer Vision RWTH Aachen University Aachen Germany Cere

International challenges have become the de facto standard for comparative assessment of image analysis algorithms given a specific task. Segmentation is so far the most widely investigated medical image processing task, but the various segmentation challenges have typically been organized in isolation, such that algorithm development was driven by the need to tackle a single specific clinical problem. We hypothesized that a method capable of performing well on multiple tasks will generalize well to a previously unseen task and potentially outperform a custom-designed solution. To investigate the hypothesis, we organized the Medical Segmentation Decathlon (MSD)—a biomedical image analysis challenge, in which algorithms compete in a multitude of both tasks and modalities. The underlying data set was designed to explore the axis of difficulties typically encountered when dealing with medical images, such as small data sets, unbalanced labels, multi-site data and small objects. The MSD challenge confirmed that algorithms with a consistent good performance on a set of tasks preserved their good average performance on a different set of previously unseen tasks. Moreover, by monitoring the MSD winner for two years, we found that this algorithm continued generalizing well to a wide range of other clinical problems, further confirming our hypothesis. Three main conclusions can be drawn from this study: (1) state-of-the-art image segmentation algorithms are mature, accurate, and generalize well when retrained on unseen tasks;(2) consistent algorithmic performance across multiple tasks is a strong surrogate of algorithmic generalizability;(3) the training of accurate AI segmentation models is now commoditized to non AI *** Codes 68T07. © 2021, CC BY-NC-SA.

关键词： Image segmentation