A simple, LabelMe-like annotation tool for AI4med data preprocessing with an intuitive GUI interface. This tool allows users to easily annotate medical images by clicking to place 600×600 pixel rectangular regions.

• Simple GUI Interface: LabelMe-like interface that's easy to use
• Click-to-Annotate: Simply click on the image to place a 600×600 pixel rectangular annotation
• Automatic Window Sizing: Adapts to your screen resolution for optimal viewing
• Scroll Support: Automatically shows horizontal/vertical scrollbars for large images
• JSON Export: Saves annotations in LabelMe-compatible JSON format
• Image Support: Supports common image formats (PNG, JPG, JPEG, BMP, TIF, TIFF)

• Python 3.6 or higher
• PyQt5

# Install PyQt5
pip install pyqt5

# Start the tool and select an image through the GUI
python main.py

# Or directly load an image
python main.py path/to/your/image.png

1. Open an Image: Click the "打开图片 (Open)" button to select an image file
2. Annotate: Click anywhere on the image to place a 600×600 pixel rectangular annotation
3. Save: Click "保存标签 (Save)" to export the annotation as a JSON file

The tool exports annotations in LabelMe-compatible JSON format:

{
  "version": "5.8.3",
  "flags": {},
  "shapes": [
    {
      "label": "d",
      "points": [[x1, y1], [x2, y2]],
      "group_id": null,
      "description": "",
      "shape_type": "rectangle",
      "flags": {},
      "mask": null
    }
  ],
  "imagePath": "image_name.jpg",
  "imageData": "base64_encoded_image_data",
  "imageHeight": 1080,
  "imageWidth": 1920
}

• Automatic Resolution Adaptation: The window automatically adjusts to your screen size
• Scroll Support: Large images automatically get scrollbars for easy navigation
• Real-time Drawing: Annotations are drawn immediately when you click
• Boundary Clipping: Rectangles are automatically clipped to image boundaries
• Base64 Image Encoding: Images are embedded in the JSON for complete annotation files

labelPZY/
├── main.py          # Main application file
├── .gitignore       # Git ignore rules (only tracks main.py)
└── d/              # Output directory for JSON files
    └── *.json      # Generated annotation files

• PyQt5: For the GUI interface
• pathlib: For file path handling
• json: For JSON export
• base64: For image encoding

This tool was developed for AI4med data preprocessing workflows, providing a simple and efficient way to annotate medical images for machine learning training.

• Added "Open Image" button functionality
• Automatic resolution adaptation
• Horizontal/vertical scrollbar support for large images
• Improved button layout

This tool is designed for AI4med data preprocessing workflows.

Feel free to submit issues and enhancement requests for this medical image annotation tool.