Segmentation of Low-Grade Glioma

This project focuses on segmenting low-grade gliomas from brain MRI scans using deep learning models.

🧠 Methodology

• Models:
  • ResNet18 – tumor presence classifier.
  • U-Net – main segmentation model (31M parameters).
  • U-Net Mini – lightweight version (1.8M parameters).
• Loss Functions:
  • BCEWithLogitsLoss (with class weighting).
  • Dice Loss for evaluation.

📊 Results

• ResNet18: 98% accuracy on tumor presence classification.
• Segmentation Performance (Dice coefficient):
  • ResNet18 + U-Net → 93.1% (all images), 79.1% (glioma only).
  • ResNet18 + U-Net Mini → 89.9% (all images), 68.9% (glioma only).
• Competitive with Kaggle solutions (best reported Dice ≈ 96%).

📖 References

• Kaggle Dataset
• Siddique et al., U-Net and Its Variants for Medical Image Segmentation, IEEE Access, 2021.