How to extract 2D and 3D features of video?

vsislab / Controllable_XGating

ICCV2019: Controllable Video Captioning with POS Sequence Guidance Based on Gated Fusion Network

67 stars 13 forks source link

Open q-rong opened 4 years ago

q-rong commented 4 years ago

Hi,

Thank you for releasing your code.

Can you tell me how you extracted 2D and 3D features from the video? As in your paper, inception_resnet-v2, I3D,C3D and so on.

Soldelli commented 4 years ago

Dear @q-rong any luck with this? I'm facing the same issue.