Hi, I have yet another question about the annotations of social_test set. In the train/val annotations, it is possible to tell which participants are in the same frame or appear in the same clip. However, the test set only gives cropped face track and corresponding audio of the clip. I'm curious if there is also a way to know which participants appear in the same scene in the test set.
SJTUwxz
commented
1 year ago
Hi, I have yet another question about the annotations of social_test set. In the train/val annotations, it is possible to tell which participants are in the same frame or appear in the same clip. However, the test set only gives cropped face track and corresponding audio of the clip. I'm curious if there is also a way to know which participants appear in the same scene in the test set.
Thank you!