Related Experiment Video
Updated: May 3, 2026
05:33
Three-Dimensional Shape Modeling and Analysis of Brain Structures
Published on: November 14, 2019
6.6K
Semi-automatic brain tumor segmentation by constrained MRFs using structural trajectories
Liang Zhao1, Wei Wu2, Jason J Corso3
1Computer Science and Engineering, SUNY at Buffalo, Buffalo, NY, USA. lzhao6@buffalo.edu
Medical Image Computing and Computer-Assisted Intervention : MICCAI ... International Conference on Medical Image Computing and Computer-Assisted Intervention
|February 8, 2014
Summary
This study introduces a semi-automatic brain tumor segmentation method for multi-channel MR images. The novel approach improves accuracy by using iterative optimization and structural trajectories, outperforming existing methods.
Area of Science:
- Medical Imaging
- Computational Biology
- Artificial Intelligence
Background:
- Accurate brain tumor quantification is crucial for prognosis.
- Fully automatic segmentation methods struggle with tumor variability.
- Clinical adoption of current methods is limited.
Purpose of the Study:
- To develop a semi-automatic segmentation framework for multi-channel MR images.
- To overcome limitations of fully automatic brain tumor segmentation.
- To improve the clinical utility of brain tumor volume and growth quantification.
Main Methods:
- A semi-automatic framework for multi-channel MR image segmentation.
- Iterative multi-label Markov Random Field (MRF) optimization with hard constraints.
- Utilizing structural trajectories for pixel correspondence between slices.
Main Results:
- Demonstrated robustness and effectiveness on the 2012 MICCAI BRATS Challenge Dataset.
- Achieved superior performance compared to baseline methods.
- Validated the utility of the constrained MRF formulation for brain tumor segmentation.
Conclusions:
- The proposed semi-automatic framework offers a viable solution for brain tumor segmentation.
- The method effectively handles tumor variability without requiring prior models.
- This approach shows promise for improved clinical prognostic measures.
