search

SeanChenxy / HandMesh

MIT License
344 stars 68 forks source link

paper issue #1

Closed qdd1234 closed 3 years ago

qdd1234 commented 3 years ago

Hi Thans for your wonderful work. I have some questions referring to 2D hand pose estimtion followed: 1.You train your model on a 3D benchmark,But Your model can still estimation 2D estimation,How do your model achieve? 2.Dose model can train on a 2D benchmark directly?

Thanks

SeanChenxy commented 3 years ago

Hi, Please refer to Figure 2.

  1. In "2D cue extraction", We use the hourglass network to predict 2D information, and training data contains 2D annotations.
  2. If you try this repo on a 2D benchmark, you would like to remove "3D mesh recovery" and corresponding loss fuctions.
qdd1234 commented 3 years ago

Could you share the code about RHD dateset?

SeanChenxy commented 3 years ago

It could come after the conference. You now can refer to https://github.com/lixiny/bihand?utm_source=catalyzex.com