D-FINE Detection Demo

This demo compares annotated ground truth data (in red) and model predictions (in blue).
Use the slider to visually compare both views:

• The left image shows the annotated labels.
• The right image displays predictions from the selected D-FINE model, with each bounding box and its confidence score.

📂 Training Dataset: All D-FINE variants were trained on the L&A Pucks Dataset available on Hugging Face.

Select a sample image below:

Citation

If you use D-FINE or its methods in your work, please cite the following BibTeX entry:

@misc{peng2024dfine,
    title={D-FINE: Redefine Regression Task in DETRs as Fine-grained Distribution Refinement},
    author={Yansong Peng and Hebei Li and Peixi Wu and Yueyi Zhang and Xiaoyan Sun and Feng Wu},
    year={2024},
    eprint={2410.13842},
    archivePrefix={arXiv},
    primaryClass={cs.CV}
}