Label format converter

If you have ever downloaded a public dataset to play with YOLO but found that the labels come in a different format, such as CVAT, COCO or PASCAL-VOC, I have created this script to make the conversion task easier.

I have tried to make it as modular as possible so that it can be integrated in other projects.

I hope you find it useful.

Dependencies

cd label-format-converter
python3 -m pip install -r requeriments.txt

Usage

Convert labels to YOLO format.

1. Place your labels into the correct folder.
   • CVAT: ~/label-format-converter/labels/cvat/*.xml
   • COCO: ~/label-format-converter/labels/coco/*.json
   • PASCAL: ~/label-format-converter/label/Annotations/*.xml
2. (CVAT/COCO) Rename labels file.
   • CVAT: annotations.xml
   • COCO: instances_default.xml
3. Update res/classes.json with the classes of the dataset
4. Launch de main script with the origin format label as argument.

cd label-format-converter
python3 main.py [-f --format] [cvat, coco, pascal]

5. Yolo folder will be created and the new labels will be placed in the next directory: ~/label-format-converter/labels/yolo

License

GNU GPL

Author

Code with 💛 by Adrián Rodríguez Galisteo

• Github: https://github.com/adrirg9412/

• LinkedIn: https://www.linkedin.com/in/adrian-rodriguez-galisteo/

• Mail: [email protected]