A medical web application powered by AI for automated analysis of brain MRI scans, specifically designed to detect and classify glioma brain tumors.

• 🔍 Automated Tumor Detection: Instantly identifies presence of brain tumors in MRI scans
• 📊 Tumor Segmentation: Precisely segments tumor regions using Attention U-Net architecture
• 🏷️ Grade Classification: Classifies tumors as HGG (aggressive) or LGG (less aggressive) using ConvNeXt
• 📄 Report Generation: Creates comprehensive PDF reports for clinical use
• 👤 User Management: Secure authentication and profile management for medical professionals
• 💾 Report Storage: Stores and retrieves patient reports with database integration

• Frontend: HTML5, CSS3, Bootstrap 5, JavaScript
• Backend: Django (Python)
• AI Models:
  • Attention U-Net for tumor segmentation
  • ConvNeXt-Base for grade classification
• Database: SQLite3
• Dataset: BraTS 2019

• Python 3.8+
• pip
• Git

1. Clone the repository

   git clone https://github.com/UsmanK7/Neuro-Insight.git
   cd neuro-insight

2. Create virtual environment

   python -m venv venv
   source venv/bin/activate  # On Windows: venv\Scripts\activate

3. Install dependencies

   pip install -r requirements.txt

4. Set up database

   python manage.py migrate

5. Run the application

   python manage.py runserver

6. Access the application

   • Open your browser and navigate to http://localhost:8000

1. Sign Up/Login: Create an account or login as a medical professional
2. Upload MRI Scans: Upload FLAIR and T1CE modalities with slice selection
3. AI Analysis: The system automatically:
   • Detects tumor presence
   • Segments tumor regions if present
   • Classifies tumor grade (HGG/LGG)
4. View Results: Review segmented images and classification results
5. Generate Reports: Create and download comprehensive diagnostic reports
6. Manage Reports: Access saved reports from your profile

• Accuracy: 99.30%
• Dice Coefficient: 99.47%
• Mean IoU: 50.00%
• Precision: 99.38%
• Sensitivity: 99.24%

• Accuracy: 98.66%
• AUC: 99.50%
• Sensitivity: 99.54%
• Precision: 94.30%
• F1-Score: 99.20%

• Source: BraTS 2019 (Brain Tumor Segmentation Challenge)
• MRI Sequences: T1CE and FLAIR modalities
• Segmentation Labels: 4 classes (background, necrotic core, edema, enhancing tumor)
• Classification Labels: HGG vs LGG

• User authentication and session management
• Secure file upload validation
• Data encryption and privacy protection
• HIPAA-compliant data handling practices

• Primary: Doctors and radiologists
• Secondary: Medical students and researchers
• Requirements: No technical AI/ML expertise needed

• ⚡ Faster Diagnosis: Reduces manual analysis time from hours to seconds
• 🎯 Improved Accuracy: AI-powered precision reduces human error
• 🌐 Enhanced Accessibility: Web-based platform for telemedicine
• 📚 Educational Value: Learning tool for medical professionals
• 💼 Workflow Integration: Seamless integration into clinical workflows

• Support for additional MRI modalities (T1, T2, T1-Gd)
• Expansion to other brain tumor types (meningioma, pituitary)
• Full 3D volume analysis
• Integration with hospital PACS systems
• Multi-language support
• Mobile application development

• BraTS 2019 dataset organizers
• University of Mianwali Computer Science Department
• Open source community for tools and libraries
• Medical professionals who provided domain expertise

⚠️ Medical Disclaimer: This application is for research and educational purposes. Always consult qualified medical professionals for clinical decisions. The AI predictions should be used as supplementary tools alongside professional medical judgment.