Related Experiment Video
Updated: Sep 12, 2025
04:48
Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
Published on: November 30, 2022
2.9K
Brain tumor segmentation by optimizing deep learning U-Net model
Abdullah A Asiri1, Lal Hussain2,3, Muhammad Irfan4
1Radiological Sciences Department, College of Applied Medical Sciences, Najran University, Najran, Kingdom of Saudi Arabia.
Summary
This study introduces a new UNet model for precise brain tumor segmentation in MRI scans. The advanced model achieves high accuracy, aiding in earlier and more effective diagnosis for improved patient outcomes.
Area of Science:
- Medical Image Analysis
- Artificial Intelligence in Healthcare
- Neuro-oncology Imaging
Background:
- Magnetic Resonance Imaging (MRI) is essential for brain tumor diagnosis.
- Accurate segmentation of complex brain tumors in MRI is challenging.
- Early tumor detection is vital for patient prognosis.
Purpose of the Study:
- To develop and assess a novel UNet-based architecture for enhanced brain tumor segmentation in MRI.
- To improve the accuracy and efficiency of brain tumor identification using deep learning.
Main Methods:
- A novel UNet architecture was designed, incorporating Leaky ReLU, batch normalization, and regularization.
- Focused loss and generalized Dice loss functions were utilized to handle class imbalance.
- The model was trained and validated on the BraTS'2020 dataset.
Main Results:
- The proposed model achieved 99.64% accuracy on the BraTS'2020 dataset.
- High Dice coefficients were obtained for necrotic core (0.8984), edema (0.8431), and enhancing tumor (0.8824).
Conclusions:
- The developed UNet-based approach demonstrates high efficacy in brain tumor segmentation.
- This method has the potential to significantly enhance diagnostic systems and patient care.
