Cloud-Based Evaluation of Anatomical Structure Segmentation and Landmark Detection Algorithms: VISCERAL Anatomy Benchmarks
Journal article, 2016
Variations in the shape and appearance of anatomical structures in medical images are often relevant radiological signs of disease. Automatic tools can help automate parts of this manual process. A cloud-based evaluation framework is presented in this paper including results of benchmarking current state-of-the-art medical imaging algorithms for anatomical structure segmentation and landmark detection: the VISCERAL Anatomy benchmarks. The algorithms are implemented in virtual machines in the cloud where participants can only access the training data and can be run privately by the benchmark administrators to objectively compare their performance in an unseen common test set. Overall, 120 computed tomography and magnetic resonance patient volumes were manually annotated to create a standard Gold Corpus containing a total of 1295 structures and 1760 landmarks. Ten participants contributed with automatic algorithms for the organ segmentation task, and three for the landmark localization task. Different algorithms obtained the best scores in the four available imaging modalities and for subsets of anatomical structures. The annotation framework, resulting data set, evaluation setup, results and performance analysis from the three VISCERAL Anatomy benchmarks are presented in this article. Both the VISCERAL data set and Silver Corpus generated with the fusion of the participant algorithms on a larger set of non-manually-annotated medical images are available to the research community.
Evaluation framework
landmark detection
organ segmentation
Author
Oscar Jimenez-Del-Toro
Hopitaux universitaires de Geneve
Haute Ecole Specialisee de Suisse occidentale
Henning Muller
Haute Ecole Specialisee de Suisse occidentale
Hopitaux universitaires de Geneve
Markus Krenn
Medical University of Vienna
Katharina Gruenberg
University Hospital Heidelberg
Abdel Aziz Taha
Vienna University of Technology
Marianne Winterstein
University Hospital Heidelberg
Ivan Eggel
Haute Ecole Specialisee de Suisse occidentale
Antonio Foncubierta-Rodriguez
Swiss Federal Institute of Technology in Zürich (ETH)
Orcun Goksel
Swiss Federal Institute of Technology in Zürich (ETH)
Andras Jakab
Medical University of Vienna
Georgios Kontokotsios
Vienna University of Technology
Georg Langs
Medical University of Vienna
Bjoern H. Menze
Swiss Federal Institute of Technology in Zürich (ETH)
Tomas Salas Fernandez
Swiss Federal Institute of Technology in Zürich (ETH)
Agency for Health Quality and Assessment of Catalonia
Roger Schaer
Haute Ecole Specialisee de Suisse occidentale
Anna Walleyo
University Hospital Heidelberg
Marc Andre Weber
University Hospital Heidelberg
Yashin Dicente Cid
Haute Ecole Specialisee de Suisse occidentale
Hopitaux universitaires de Geneve
Tobias Gass
Swiss Federal Institute of Technology in Zürich (ETH)
Mattias Heinrich
Universitaet Zu Lübeck
Fucang Jia
Chinese Academy of Sciences
Fredrik Kahl
Chalmers, Signals and Systems, Signal Processing and Biomedical Engineering
Razmig Kechichian
Université de Lyon
Dominic Mai
University of Freiburg
Assaf B. Spanier
The Hebrew University Of Jerusalem
Graham Vincent
Imorphics Ltd
C. L. Wang
Royal Institute of Technology (KTH)
Daniel Wyeth
Toshiba Medical Visualization Systems Europe
Allan Hanbury
Vienna University of Technology
IEEE Transactions on Medical Imaging
0278-0062 (ISSN) 1558254x (eISSN)
Vol. 35 11 2459-2475 7488206Subject Categories
Signal Processing
DOI
10.1109/TMI.2016.2578680