search

yfeng95 / PRNet

Joint 3D Face Reconstruction and Dense Alignment with Position Map Regression Network (ECCV 2018)
http://openaccess.thecvf.com/content_ECCV_2018/papers/Yao_Feng_Joint_3D_Face_ECCV_2018_paper.pdf
MIT License
4.95k stars 947 forks source link

wrong result of the output of estimate_pose #169

Open Guocode opened 4 years ago

Guocode commented 4 years ago

I draw the pose(yaw pitch roll) which is the second parameter of estimate_pose, it is right when yaw is small but it is wrong when yaw is large, here is the example

20200224-163017(eSpace)

My opinion is that the parameter order should be like this compute_similarity_transform(canonical_vertices,vertices)

smonefrmsouth commented 4 years ago

Hi Team, I tried running the code for pose estimation but I get values in between less than 1. Is the output scaled or is the model prediction wrong values?

lyyiangang commented 4 years ago

Me too. I think we should reverse the input parameters of compute_similarity_transform. compute_similarity_transform(canonical_vertices,vertices) should be reasonable.