FreeDA: Training-Free Open-Vocabulary Segmentation with Offline Diffusion-Augmented Prototype Generation (CVPR 2024)

This is the repository that contains source code for the FreeDA website.

If you find FreeDA useful for your work please cite:

@article{barsellotti2024training
  title={Training-Free Open-Vocabulary Segmentation with Offline Diffusion-Augmented Prototype Generation},
  author={Barsellotti, Luca and Amoroso, Roberto and Cornia, Marcella and Baraldi, Lorenzo and Cucchiara, Rita},
  booktitle={Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition},
  year={2024}
}