@article{hmenze2015the, author = {H. Menze, Bjoern and Jakab, Andras and Bauer, Stefan and Kalpathy-Cramer, Jayashree and Farahani, Keyvan and Kirby, Justin and Burren, Yuliya and Porz, Nicole and Slotboom, Johannes and Wiest, Roland and Lanczi, Levente and Gerstner, Elizabeth and Weber, Marc-Andre and Arbel, Tal and B. Avants, Brian and Ayache, Nicholas and Buendia, Patricia and Louis Collins, D. and Cordier, Nicolas and J. Corso, Jason and Criminisi, Antonio and Das, Tilak and Delingette, Hervé and Demiralp, Cagatay and R. Durst, Christopher and Dojat, Michel and Doyle, Senan and Festa, Joana and Forbes, Florence and Geremia, Ezequiel and Glocker, Ben and Golland, Polina and Guo, Xiaotao and Hamamci, Andac and M. Iftekharuddin, Khan and Jena, Raj and M. John, Nigel and Konukoglu, Ender and Lashkari, Danial and Antonio Mariz, Jose and Meier, Raphael and Pereira, Sergio and Precup, Doina and J. Price, Stephen and Riklin Raviv, Tammy and M. S. Reza, Syed and Ryan, Michael and Sarikaya, Duygu and Schwartz, Lawrence and Shin, Hoo-Chang and Shotton, Jamie and A. Silva, Carlos and Sousa, Nuno and K. Subbanna, Nagesh and Szekely, Gabor and J. Taylor, Thomas and M. Thomas, Owen and J. Tustison, Nicholas and Unal, Gozde and Vasseur, Flor and Wintermark, Max and Hye Ye, Dong and Zhao, Liang and Zhao, Binsheng and Zikic, D. and Prastawa, Marcel and Reyes, Mauricio and Van Leemput, Koen}, title = {The Multimodal Brain Tumor Image Segmentation Benchmark (BRATS)}, year = {2015}, month = {September}, abstract = {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.}, url = {http://approjects.co.za/?big=en-us/research/publication/the-multimodal-brain-tumor-image-segmentation-benchmark-brats/}, pages = {1993-2024}, journal = {IEEE Transactions on Medical Imaging}, volume = {34}, number = {10}, }