Classification stage and grounding stage

[guanw-pku]

Hi, I have a one question.

Could you please explain the choice that only applying classification stage on VidVRD dataset?

Thank you!

[Dawn-LX]

Hi, I have a one question.

Could you please explain the choice that only applying classification stage on VidVRD dataset?

Thank you!

refer to Section 4.2 of our paper

. For each relation triplet in the training set of VidVRD [35], we noticed that only a portion of ground-truth segments is annotated as foreground, which makes the annotated temporal boundaries unreliable for training. Therefore, we only used the classification stage for VidVRD

[guanw-pku]

Thank you!