Questions about UVO Dataset

[lei12879]

Great job! I have some questions about the UVO dataset.

Did you directly use the official 'video2frames.py' script from the website to convert videos into images?

Because the paper mentions that the UVO dataset has 15k images, but the official dataset contains 10k videos.

So, I would like to know if only 50 videos were processed. I used the official 'video2frames.py' script to process 50 videos, but the resulting files do not match the annotation file.

Looking forward to your response!