Are there duplicate identities across the 330K clips?
#4
by
KimRRRich - opened
This dataset is a valuable contribution to the video generation community — having 330K videos at this scale is a significant resource for advancing research in human-centric video synthesis and generation.
Thank you for sharing it!
I'd like to ask: are there any duplicate individuals (subjects/persons) appearing across multiple videos in this dataset?
For example:
Does the same person appear in more than one video clip?
Was any de-duplication process applied at the identity level during data collection or preprocessing?
This would be very helpful
Q1: Does the same person appear in more than one video clip?
A1: Yes
Q2: Was any de-duplication process applied at the identity level during data collection or preprocessing?
A2: Yes