Closed tjusxh closed 5 years ago
Thanks for your big contribution. I have a question, when you training the network, how to get the ground truth p corresponding for the image? My guess: you got them by fitting 3dmm. is it right?
The ground-truth is from 300W-LP dataset. You can also generate them by 3DMM fitting as you said.
Thanks for your big contribution. I have a question, when you training the network, how to get the ground truth p corresponding for the image? My guess: you got them by fitting 3dmm. is it right?