Nodule Detection API

This application provides a Flask-based REST API for detecting nodules in medical images using a UNet model.

Project Structure

nodule_detection/
├── app.py                      # Main application entry point
├── config.py                   # Configuration settings
├── utils/
│   ├── __init__.py
│   ├── logging_config.py       # Logging configuration
│   └── image_utils.py          # Image processing utilities
├── model/
│   ├── __init__.py
│   ├── loader.py               # Model loading functions
│   └── inference.py            # Model inference functions
├── features/
│   ├── __init__.py
│   ├── basic_features.py       # Basic feature calculations
│   └── radiomic_features.py    # PyRadiomics feature extraction
├── visualization/
│   ├── __init__.py
│   └── visualizer.py           # Visualization functions
└── api/
    ├── __init__.py
    ├── routes.py               # API endpoints
    └── response_utils.py       # Helper functions for API responses

Key Features

• Medical image nodule detection using UNet deep learning model
• Remote model downloading capability
• Processing of multiple images in batch
• Calculation of radiomic features for detected nodules
• Visualization of detection results with bounding boxes
• Progress tracking for long-running operations

API Endpoints

• /health - Check service health and model status
• /load_model - Explicitly load or check model loading status
• /progress - Get current processing progress
• /predict_multiple - Process multiple images and return visualizations with nodule data

Requirements

• Python 3.6+
• PyTorch
• Flask
• SimpleITK
• PyRadiomics
• Matplotlib
• NumPy
• SciPy

Installation

1. Clone the repository
2. Install required packages: pip install -r requirements.txt
3. Run the application: python app.py