TrippleD

Train Datas include "image.png + landmarks.npy(FAN detect) + segmentation.npy(face_segmentation detect)"

1. face_segmentation part. You must install caffe(with gpu).

The follow picture should be made by Vggface2_train_list_max_normal_100_ring_5_1serial.npy each candidate : image/225/0000383.png, 68kpts/225/0000383.npy, face_segmental/225/0000383.npy,  [shape_consistent_loss & detail_loss] require images >= 3 for each candidate

Maybe can use 'self.model = face_alignment.FaceAlignment(1, flip_input=False)' instead

I think you can you FLAME_w_HIFI3D_UV.obj, tex_uv.png instead flame model. And you should use hifi3dface tool to generate UV basis.

> Were you able to resolve this? No. If you can train well, please tell me.


@csbhr


Hi , bro. I find that the tiago_lar.bag does not exist. Can you provide a new data set connection? My email is [email protected]. Thanks a lot!