search

MichiganCOG / Video-Grounding-from-Text

Source code for "Weakly-Supervised Video Object Grounding from Text by Loss Weighting and Object Interaction"
Other
44 stars 9 forks source link

How to know related frame of given bounding box annotations #8

Open zeynepgokce opened 2 years ago

zeynepgokce commented 2 years ago

Hi, I want to learn more information about bounding box annotations in download page of Youcook2 dataset. There is a json file that contains the bounding box annotations of the related object for validation set. But I could not obtain related frame number of given bounding boxes. Can I obtain ground truth bounding boxes with frame of video segments ?

Not: It is not about faster-RCNN predicted top 100 bounding boxes. It is about the ground truth object bounding box annotations.

Thank you so much

zeynepgokce commented 2 years ago

I guess, the frames correspond to frames of related segment which are extracted in fps=1.