29-08-2016, 03:35 PM
Proceedings of the MICCAI Challenge on Multimodal
Brain Tumor Image Segmentation (BRATS) 2013
1451487399-procbrats2013.pdf (Size: 2.56 MB / Downloads: 11)
Preface
Because of their unpredictable appearance and shape, segmenting brain tumors
from multi-modal imaging data is one of the most challenging tasks in medical
image analysis. Although many different segmentation strategies have been
proposed in the literature, it is hard to compare existing methods because the
validation datasets that are used differ widely in terms of input data (structural
MR contrasts; perfusion or diffusion data; ...), the type of lesion (primary or
secondary tumors; solid or infiltratively growing), and the state of the disease
(pre- or post-treatment).
In order to gauge the current state-of-the-art in automated brain tumor
segmentation and compare between different methods, we are organizing a
Multimodal Brain Tumor Segmentation (BRATS) challenge that is held in
conjunction with the 16
th International Conference on Medical Image Computing
and Computer Assisted Intervention (MICCAI 2013) on September 22nd, 2013 in
Nagoya, Japan. This event succeeds the MICCAI-BRATS 2012 challenge that was
held in conjunction with MICCAI 2012.
For this purpose, we are making available a large dataset of brain tumor MR
scans in which the tumor and edema regions have been manually delineated,
adding another 20 multimodal image volume from high and low grade glioma
patients to the BRATS 2012 data set. All images – in both the publicly
distributed training data set, and the blinded test data set- are annotated
through clinical experts who annotated four different types of tumor substructurs
(edema, enhancing core, non-enhancing core, necrotic core).
Participating teams downloaded the training data for algorithmic tweaking and
tuning. The teams then evaluated their segmentation performance on the
training data, and submitted a short paper describing the results and the
segmentation method that was used that were subsequently reviewed by the
organizing committee. A total of 10 submissions were accepted for the final
challenge. The corresponding short papers describing training results and the
methodological approaches are summarized in this volume. On the day of the
challenge itself, an independent set of test scans is made available and analyzed
on the spot by each team, after which the methods are ranked according to their
performance.