About grounding dino

[xiang-xiang-zhu]

I ran the grounding dino using the following shell

python demo/image_demo.py demo/demo.jpg \
        configs/mm_grounding_dino/grounding_dino_swin-t_pretrain_obj365.py \
        --weights grounding_dino_swin-t_pretrain_obj365_goldg_grit9m_v3det_20231204_095047-b448804b.pth \
        --texts 'chair' -c

Then in the output folder, a demo.json file and a visualization image file are generated. In the demo.json is a dict containing 'labels' and 'scores', but I don't know what they mean respectively?
If I want to get the bounding box of the object detection, what should I do, thanks!