search

hitachinsk / FGT

[ECCV 2022] Flow-Guided Transformer for Video Inpainting
https://hitachinsk.github.io/publication/2022-10-01-Flow-Guided-Transformer-for-Video-Inpainting
MIT License
311 stars 32 forks source link

The flow propagation #35

Closed diaodeyi closed 1 year ago

diaodeyi commented 1 year ago

Hi , thanks for great work! The flow propagation mentioned in paper seems to be missing in the released code, it doesn't matter?

hitachinsk commented 1 year ago

Our repository contains the flow-guided content propagation. Please refer https://github.com/hitachinsk/FGT/blob/master/tool/get_flowNN_gradient.py for more details.

diaodeyi commented 1 year ago

The flow-guided content propagation is just for inference. In this way, the input setting is different for training and inference?

hitachinsk commented 1 year ago

Yes, the flow-guided content propagation is just for inference, and there is slight difference between the training and the inference procedure.