Related Experiment Video
Updated: Sep 28, 2025
10:48
PET and MRI Guided Irradiation of a Glioblastoma Rat Model Using a Micro-irradiator
Published on: December 28, 2017
9.6K
Category guided attention network for brain tumor segmentation in MRI
Jiangyun Li1, Hong Yu1, Chen Chen2
1School of Automation and Electrical Engineering, University of Science and Technology Beijing, Beijing 100083, People's Republic of China.
Physics in Medicine and Biology
|March 30, 2022
Summary
This study introduces the Category Guided Attention U-Net (CGA U-Net) for improved brain tumor segmentation in MRI scans. The novel network enhances accuracy and reduces computational cost, aiding radiation treatment planning.
Area of Science:
- Medical Imaging
- Artificial Intelligence
- Neuroscience
Background:
- Magnetic resonance imaging (MRI) is crucial for brain disease diagnosis.
- Accurate brain tumor segmentation is vital for effective radiation therapy.
- Low tissue contrast in tumors presents a significant segmentation challenge.
Purpose of the Study:
- To develop a novel network for accurate and automatic brain tumor segmentation.
- To address the challenge of low tissue contrast in MRI tumor regions.
- To improve segmentation performance and reduce computational complexity.
Main Methods:
- Proposed a Category Guided Attention U-Net (CGA U-Net).
- Introduced a Supervised Attention Module (SAM) for enhanced feature map dependency.
- Implemented an intra-class update approach for feature map reconstruction.
Main Results:
- The CGA U-Net outperformed state-of-the-art algorithms on BraTS 2019 datasets.
- Achieved superior segmentation performance.
- Demonstrated reduced computational complexity compared to existing methods.
Conclusions:
- The CGA U-Net effectively captures global semantic information using SAM.
- Significantly reduces computational cost in brain tumor segmentation.
- Offers a promising solution for improving radiation treatment planning.
