Related Experiment Video
Updated: May 8, 2026
06:48
Automated Segmentation of Cortical Grey Matter from T1-Weighted MRI Images
Published on: January 7, 2019
Ensemble segmentation for GBM brain tumors on MR images using confidence-based averaging
Jing Huo1, Kazunori Okada, Eva M van Rikxoort
1TeraRecon Inc., 4000 East 3rd Avenue, Suite 200, Foster City, California 94404, USA. jinghuo@gmail.com
Medical Physics
|September 7, 2013
Summary
This study developed an ensemble segmentation framework for glioblastoma multiforme tumors using confidence map averaging (CMA). The CMA ensemble achieved results comparable to the best individual segmentation methods.
Area of Science:
- Medical imaging analysis
- Artificial intelligence in oncology
Background:
- Ensemble methods enhance segmentation accuracy and robustness.
- Accurate glioblastoma multiforme segmentation is crucial for treatment planning.
Purpose of the Study:
- Develop an ensemble segmentation framework for glioblastoma multiforme tumors.
- Utilize single-channel T1w postcontrast magnetic resonance images.
Main Methods:
- Evaluated three base segmentation methods: fuzzy connectedness, GrowCut, and support vector machine voxel classification.
- Employed confidence map averaging (CMA) as the ensemble rule.
Main Results:
- Performance assessed on a dataset of 46 glioblastoma cases.
- Segmentation accuracy evaluated using the F1-measure against ground truth.
Conclusions:
- The CMA ensemble method statistically approximated the best individual segmentation results.
- Demonstrated the efficacy of ensemble learning for glioblastoma segmentation.
