The Multimodal Brain Tumor Image Segmentation Benchmark (BRATS)

Bjoern Menze 1 Andras Jakab 2 Stefan Bauer 3 Jayashree Kalpathy-Cramer 4 Keyvan Farahani 5 Justin Kirby 5 Yuliya Burren 6 Nicole Porz 6 Johannes Slotboom 6 Roland Wiest 6 Levente Lanczi 7 Elisabeth Gerstner 4 Marc-Andre Weber 8, 9 Tal Arbel 10 Brian Avants 11 Nicholas Ayache 1 Patricia Buendia 12 Louis Collins 13 Nicolas Cordier 1 Jason Corso 14 Antonio Criminisi 15 Tilak Das 16 Hervé Delingette 1 Cagatay Demiralp 17 Christopher Durst 18 Michel Dojat 19 Senan Doyle 20 Joana Festa 21 Florence Forbes 20 Ezequiel Geremia 1 Ben Glocker 15 Polina Golland 22 Xiaotao Guo 23 Andac Hamamci 24 Khan Iftekharuddin 25 Raj Jena 26 Nigel John 27 Ender Konukoglu 4 Danial Lashkari 22 Jose Antonio Mariz 21 Raphael Meier 28 Sergio Pereira 29 Doina Precup 30 S. J. Price 31 Tammy Riklin-Raviv 22 Syed Reza 25 Michael Ryan 12 Lawrence Schwartz 23 Hoo-Chang Shin 32 Jamie Shotton 15 Carlos Silva 21 Nuno Sousa 33 Nagesh Subbanna 30 Gabor Szekely 34 Thomas Taylor 12 Owen Thomas 12 Nicholas Tustison 18 Gozde Unal 35 Flor Vasseur 19 Max Wintermark 36 Dong Hye Ye 37 Liang Zhao 14 Binsheng Zhao 38 Darko Zikic 15 Marcel Prastawa 39 Mauricio Reyes 3 Koen Van Leemput 22, 40
1 ASCLEPIOS - Analysis and Simulation of Biomedical Images
CRISAM - Inria Sophia Antipolis - Méditerranée
20 MISTIS - Modelling and Inference of Complex and Structured Stochastic Systems
Inria Grenoble - Rhône-Alpes, Institut National Polytechnique de Grenoble (INPG), LJK - Laboratoire Jean Kuntzmann
Abstract : In this paper we report the set-up and results of the Multimodal Brain Tumor Image Segmentation (BRATS) benchmark 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 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 subregions 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.
Document type :
Journal articles
IEEE Transactions on Medical Imaging, Institute of Electrical and Electronics Engineers (IEEE), 2014, pp.33. <10.1109/TMI.2014.2377694>


https://hal.inria.fr/hal-00935640
Contributor : Bjoern Menze <>
Submitted on : Saturday, November 29, 2014 - 1:13:42 PM
Last modification on : Thursday, July 9, 2015 - 2:42:01 PM

File

menze_brats.pdf
fileSource_public_author

Identifiers

Collections

Citation

Bjoern Menze, Andras Jakab, Stefan Bauer, Jayashree Kalpathy-Cramer, Keyvan Farahani, et al.. The Multimodal Brain Tumor Image Segmentation Benchmark (BRATS). IEEE Transactions on Medical Imaging, Institute of Electrical and Electronics Engineers (IEEE), 2014, pp.33. <10.1109/TMI.2014.2377694>. <hal-00935640v2>

Export

Share

Metrics

Consultation de
la notice

1565

Téléchargement du document

512