search

tsing90 / pytorch_semantic_human_matting

This is an unofficial implementation of the paper "Semantic human matting":
https://arxiv.org/pdf/1809.01354.pdf
83 stars 18 forks source link

compositing on the fly #9

Closed Oseira closed 5 years ago

Oseira commented 5 years ago

Thank you for the great work. I am trying to understand how this project works: If I understand correctly, the model is trained using 2 lists: one with foregrounds and one with backgrounds that are then composited on the fly during the training.

I can't seem to understand where in the train.py code does the compositing happen? Thanks for your help

kkkmax commented 5 years ago

According to the design of tsing90, I think that synthesis and then training, synthesis should be completed in the data preparation process

tsing90 commented 5 years ago

@kkkmax is correct, that's done in the dataloader, plz refer to the file ./data/dataset.py

Oseira commented 5 years ago

Thank you both, I think I see how it goes now.