Related Experiment Video
Updated: Jun 10, 2026
10:25
Brain Infarct Segmentation and Registration on MRI or CT for Lesion-symptom Mapping
Published on: September 25, 2019
Complementary cross-gated fusion framework for brain tumor segmentation using MR images (CoGFu Net)
1Department of Electronics and Communication Engineering, Mepco Schlenk Engineering College, Sivakasi, Tamil Nadu 626 005, India.
Journal of Neuroscience Methods
|June 8, 2026
Summary
This study introduces a novel model for brain tumor segmentation in MR images, improving accuracy by integrating contextual and structural features. The model demonstrates robust performance on benchmark datasets, aiding clinical diagnosis.
Area of Science:
- Medical Imaging
- Artificial Intelligence
- Computational Biology
Background:
- Brain tumor segmentation from MR images is crucial for diagnosis and treatment planning.
- Challenges include diverse intensity patterns and overlapping boundaries.
- Existing methods often neglect structural boundary details and inter-feature learning.
Purpose of the Study:
- To develop a novel model for accurate brain tumor segmentation.
- To address limitations of existing methods by incorporating complementary features.
- To facilitate subregion-aware prediction and clinical deployment.
Main Methods:
- A novel model combining complementary feature learning (contextual semantic and fine edge learning) with cross-gated fusion.
- Parallel processes extract global semantic context and local edge-aware details.
- Cross-gated fusion adaptively balances semantic richness and structural precision.
Main Results:
- Achieved high performance on BraTS2020 dataset: mean Dice Coefficients (DC) of 0.9489 (WT), 0.8581 (TC), 0.8352 (ET).
- Achieved mean Intersection over Union (IOU) of 0.8961 (WT), 0.8572 (TC), 0.8093 (ET) on BraTS2020.
- Demonstrated improved DC by 3.3-5.7% compared to existing methods across multiple datasets.
Conclusions:
- The proposed framework effectively segments multi-class and single-tumor regions.
- It addresses challenges in brain tumor segmentation.
- Facilitates practical subregion-aware clinical deployment.
