The Multimodal Brain Tumor Image Segmentation Benchmark (BRATS)
Бенчмарк мультимодальной сегментации изображений опухолей головного мозга (BRATS)
2014-12-04
SCID: 54.1/vp5g3fzh
Discuss with AI
BRATS benchmarkDice scorebrain tumor segmentationgliomamultimodal MR imaging
Figures from the paper
Abstract (AI)
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.
Key Findings
1
BRATS released the imaging data and manual annotations through an online evaluation system as an ongoing public benchmarking resource.
2
Different algorithms performed best on different tumor sub-regions, sometimes reaching human inter-rater variability, but none ranked among the top methods for all regions.
3
Hierarchical majority-vote fusion of several strong algorithms consistently outperformed every individual algorithm across evaluations.
4
Human raters showed substantial disagreement across tumor sub-regions, with Dice scores ranging from 74% to 85%, highlighting task difficulty.
5
The BRATS benchmark evaluated 20 state-of-the-art tumor segmentation algorithms on 65 real and 65 simulated multi-contrast MR scans.
Research Object
Multimodal Brain Tumor Image Segmentation Benchmark (BRATS) dataset and evaluation framework (multi-contrast MR scans of low- and high-grade glioma patients, simulated scans, and manual annotations)
Research Subject
accuracy, inter-rater variability, and comparative performance of tumor sub-region segmentation algorithms and their hierarchical majority-vote fusion
Publication Details
Publication Date
2014-12-04
Journal
Publisher
ISSN
Cited by
6823
Open access PDF
Access Type
Author Information
Download PDF
Subscribe to digest
References available in scid.ai1
Cited by6
Deep Convolutional Neural Networks for Computer-Aided Detection: CNN Architectures, Dataset Characteristics and Transfer Learning2016
U-Net and Its Variants for Medical Image Segmentation: A Review of Theory and Applications2021
Transformers in medical image analysis2022
Artificial intelligence and machine learning for medical imaging: A technology review2021
Federated learning enables big data for rare cancer boundary detection2022
Synthetic data in generalizable, learning-based neuroimaging2024