Hello authors,
In the paper, you mentioned that the scenes are categorized into 15 classes and that there are manually annotated region labels.
However, I could not find the corresponding category or region annotation information in the dataset released on Hugging Face.
May I ask whether these annotations will be provided in a future release, or if there is another way to access them?
Thank you very much for your help.