A modern, user-friendly desktop application for creating bounding box annotations on images. Export your annotations in COCO, YOLO, and Pascal VOC formats with a single click.

• 🖼️ Intuitive Interface: Clean, modern GUI with professional styling

• 🎯 Multi-format Export: Export annotations to COCO JSON, YOLO TXT, and Pascal VOC XML formats

• ⌨️ Keyboard Shortcuts: Navigate and annotate efficiently with keyboard controls

• 📊 Real-time Statistics: Track progress with live annotation statistics

• 🎨 Visual Feedback: Color-coded bounding boxes with class labels

• 🔄 Easy Navigation: Quickly move between images with arrow keys or buttons

• 🗂️ Class Management: Smart class selection with auto-suggestions from previous annotations

Python 3.7+

Install the required packages:

pip install -r requirements.txt

Running the Application

python main.py

Load Images: Click "Load Images" and select a folder containing your images

Draw Annotations: Click and drag to create bounding boxes around objects

Classify Objects: Enter or select a class name for each annotation

Navigate: Use arrow keys or buttons to move between images

Export: Click "Export All" to save annotations in all supported formats

• Left Arrow: Previous image
• Right Arrow: Next image
• Delete: Remove last annotation
• Ctrl+O: Load new folder
• Ctrl+S: Export annotations

COCO: JSON format with category definitions and bounding box coordinates

YOLO: TXT files with normalized coordinates and class IDs

Pascal VOC: XML files with object metadata and bounding boxes

image-annotator/
├── main.py          # Main application GUI
├── writers.py       # Export functions for different formats
├── utils.py         # Utility functions
├── requirements.txt         # Libraries needed to install
├── LICENSE/     # License
└── README.md        # This file

After export, your annotations will be organized as:

your-image-folder/
└── annotations/
    ├── coco/
    │   ├── images/
    │   └── annotations.coco.json
    ├── yolo/
    │   ├── images/
    │   ├── classes.txt
    │   └── [image_name].txt files
    └── pascal/
        ├── images/
        └── [image_name].xml files

Contributions are welcome! Please feel free to submit a Pull Request.

This project is open source and available under the MIT License.