GP-Tea: Artificial Intelligence in Healthcare

A smart healthcare companion mobile application designed to bridge the gap between patients and preliminary medical diagnosis using Artificial Intelligence.

⚡ GenAI Impact: Accelerating Development

This project demonstrates the power of AI-Assisted Engineering. By leveraging tools like Claude 3.5 Sonnet and GPT-4 for code generation and architectural guidance, we achieved a 60% reduction in R&D time.

> Comparative analysis showing over 30 hours of development time saved across Research, Coding, and Testing phases.

👩‍💻 My Contribution (Frontend & Integration)

• UI Development: Built the complete mobile interface using Flutter.
• System Integration: Connected the FastAPI backend services with the mobile frontend.
• Feature Assembly: Orchestrated various AI modules into a cohesive application. Challenge: Artificial Intelligence in Healthcare

A comprehensive healthcare application built with Flutter frontend and FastAPI backends, featuring AI-powered medical analysis tools including symptom checking, medical image classification, and lab report analysis.

👥 Team Information

Name	Role	Email
Seif-Eldeen Mostafa	AI Engineer	seifeldeen.320240021@ejust.edu.eg
Mohamed Ahmed	Back End Engineer	Mohamed.320240078@ejust.edu.eg
Sara Youssef	Front End Engineer	sara.320240074@ejust.edu.eg
Arwa Waleed	Academic Writing	arwa.320240128@ejust.edu.eg
Judy Kamal	Academic Writing	Joudy.320240043@ejust.edu.eg

Demo Video

https://drive.google.com/file/d/1d5KCNWoKFrYilVkajv_WNTjq-wIuKZGT/view?usp=sharing

📲 Try the Application

You can download and try the latest build of the AI in Healthcare application from the link below:

Download Application

🚀 Project Features Overview

🔍 AI-Powered Symptom Checker

• Intelligent Disease Prediction: XGBoost machine learning model for accurate symptom analysis
• Interactive Symptom Selection: Search and select from 132+ medical symptoms with intuitive UI
• Confidence Scoring: Ranked disease predictions with detailed confidence percentages
• Real-time Analysis: Fast processing with comprehensive medical insights

🖼️ Medical Image Classification

• Cancer Detection: Advanced TensorFlow EfficientNetV2S model for medical image analysis
• Multi-source Input: Camera capture and gallery upload support
• Instant Results: Real-time image processing with detailed confidence scores
• Professional Analysis: Structured prediction results with medical interpretations

🧪 Lab Report Analysis

• AI-Powered Document Analysis: Hugging Face inference for intelligent lab report interpretation
• Document Processing: Support for various lab report formats and images
• Structured Results: Organized analysis with summary, key findings, and medical interpretations
• Comprehensive Insights: Detailed explanations and healthcare recommendations

🛠️ Local Development Setup

Prerequisites

Before setting up the project, ensure you have the following installed:

• Flutter SDK: Version 3.0 or higher
• Python: Version 3.8 or higher
• Git: Latest version
• Android Studio/VS Code: For development
• Python Package Manager: pip

📱 Frontend Setup (Flutter Application)

1. Clone the Repository

   git clone https://github.com/Tantawi65/Healthcare.git
   cd Healthcare

2. Install Flutter Dependencies

   flutter pub get

3. Configure Flutter Environment

   flutter doctor

   Ensure all requirements are met before proceeding.

4. Run the Flutter Application

   # For Android
   flutter run
   
   # For iOS (macOS only)
   flutter run -d ios
   
   # For Web
   flutter run -d web

🐍 Backend Services Setup

The application consists of three separate FastAPI backend services that need to be started individually.

1. Symptom Checker Service (Port 8002)

# Navigate to the service directory
cd backend/Text_classification/symptom_checker

# Create and activate virtual environment
python -m venv venv

# Windows
venv\Scripts\activate
# macOS/Linux
source venv/bin/activate

# Install required dependencies
pip install -r requirements.txt

# Start the FastAPI server
uvicorn main:app --host 0.0.0.0 --port 8002 --reload

2. Image Classification Service (Port 8000)

# Navigate to the service directory
cd backend/Image_classification

# Create and activate virtual environment
python -m venv venv

# Windows
venv\Scripts\activate
# macOS/Linux
source venv/bin/activate

# Install required dependencies
pip install -r requirements.txt

# Start the FastAPI server
uvicorn main:app --host 0.0.0.0 --port 8000 --reload

3. Lab Analysis Service (Port 8003)

# Navigate to the service directory
cd backend/Lab_analysis

# Create and activate virtual environment
python -m venv venv

# Windows
venv\Scripts\activate
# macOS/Linux
source venv/bin/activate

# Install required dependencies
pip install -r requirements.txt

# Set up environment variables
cp .env.example .env
# Edit .env file and add your Hugging Face API key:
# HUGGINGFACE_API_KEY=your-api-key-here

# Start the FastAPI server
uvicorn main:app --host 0.0.0.0 --port 8003 --reload

🔧 Environment Configuration

Required Environment Variables

Create a .env file in each backend service directory with the following variables:

# Hugging Face API Key for Lab Analysis
HUGGINGFACE_API_KEY=your-huggingface-api-key-here

# API Configuration
SYMPTOM_CHECKER_PORT=8002
IMAGE_CLASSIFICATION_PORT=8000
LAB_ANALYSIS_PORT=8003

# CORS Origins (for production)
CORS_ORIGINS=http://localhost:3000,https://yourdomain.com

Flutter API Configuration

Update the API endpoints in lib/services/api_service.dart if needed:

class ApiService {
  static const String symptomCheckerBaseUrl = 'http://localhost:8002';
  static const String imageClassificationBaseUrl = 'http://localhost:8000';
  static const String labAnalysisBaseUrl = 'http://localhost:8003';
}

📦 Dependencies

Flutter Dependencies

• flutter: SDK framework
• http: API communication
• image_picker: Camera and gallery access
• go_router: Navigation management
• provider: State management

Python Backend Dependencies

• fastapi: Web framework
• uvicorn: ASGI server
• xgboost: Machine learning (Symptom Checker)
• tensorflow: Deep learning (Image Classification)
• huggingface-hub: AI inference (Lab Analysis)
• python-multipart: File upload handling
• python-jose: JWT token handling
• pillow: Image processing

🧪 Testing the Application

Verify Backend Services

1. Health Checks:

   • Symptom Checker: http://localhost:8002/health
   • Image Classification: http://localhost:8000/health
   • Lab Analysis: http://localhost:8003/health

2. API Documentation:

   • Symptom Checker: http://localhost:8002/docs
   • Image Classification: http://localhost:8000/docs
   • Lab Analysis: http://localhost:8003/docs

Test API Endpoints

# Test Symptom Checker
curl -X POST "http://localhost:8002/api/check-symptoms" \
     -H "Content-Type: application/json" \
     -d '{"symptoms": ["fever", "cough", "headache"]}'

# Test Image Classification
curl -X POST "http://localhost:8000/api/classify-image" \
     -F "image=@/path/to/medical/image.jpg"

# Test Lab Analysis
curl -X POST "http://localhost:8003/api/analyze-lab" \
     -F "image=@/path/to/lab/report.pdf"

🚀 Building for Production

Flutter Build Commands

# Android APK
flutter build apk --release

# Android App Bundle
flutter build appbundle --release

# iOS (macOS only)
flutter build ios --release

# Web
flutter build web --release

Backend Deployment

For production deployment, consider using:

• Docker containers for each service
• Environment variable management
• Load balancing for high availability
• API rate limiting and authentication
• Database integration for user data persistence

⚠️ Important Notes

• Model Files: AI model files are excluded from the repository due to size constraints. You'll need to train your own models or obtain pre-trained models separately.
• API Keys: Never commit API keys to version control. Use environment variables for all sensitive configuration.
• Security: This application is for educational and research purposes. Implement proper authentication and security measures for production use.

📄 Medical Disclaimer

This application is developed for educational and research purposes only. It should not be used as a substitute for professional medical advice, diagnosis, or treatment. Always consult with qualified healthcare professionals for medical concerns.

📲 Try the Application

You can download and try the latest build of the AI in Healthcare application from the link below:

Download Application

---

🌐 Live Deployment

The backend services are deployed and accessible online for testing and demonstration purposes.
Each service runs on a separate FastAPI instance with integrated AI models.

🔬 Lab Analysis Service

Analyze lab reports using AI-powered document understanding: 👉 https://tantawi-lab-analyzer.hf.space/

🖼️ Image Classification Service

Classify medical images and detect cancerous patterns: 👉 https://tantawi-image-class.hf.space/docs

💬 Symptom Checker (Text Classification)

Predict possible diseases based on user-entered symptoms: 👉 https://tantawi-text-classification.hf.space/docs

🧠 AI Usage

This project integrates Generative AI tools throughout the development process to enhance productivity, code quality, and documentation.
All screenshots, prompts, and AI responses used during development are organized and available in the repository under the folder "AI Usage".

📞 Support

For technical support or project inquiries, please contact any team member using the email addresses provided above.

Built with ❤️ by the GP-Tea Team for advancing AI in Healthcare