HisRepItself icon indicating copy to clipboard operation
HisRepItself copied to clipboard

Published 20 hours ago • verified publisher icon wei-mao-2019

AMASS Evaluation

Open Tim-Salzmann opened this issue 4 years ago • 3 comments

Hi,

in your paper you state that you are using the BMLrub dataset as test dataset. However in your code it looks to me like BioMotionLab_NTroje is hard coded for test. Could you provide some clarification here? https://github.com/wei-mao-2019/HisRepItself/blob/55a4186fd962388f25f21afebfbabdb493fd3451/utils/amass3d.py#L38

Also I was wondering if there is any particular reason why you are using your own forward kinematic instead of using the forward function of the Body Prior Model which returns the position for a given pose?

https://github.com/nghorbani/human_body_prior/blob/1936f38aec4bb959f6a8bf4ed304b6aafb42fa30/human_body_prior/body_model/body_model.py#L194

Thanks in advance!

Tim-Salzmann avatar Feb 09 '21 10:02 Tim-Salzmann

Just realised that "BioMotionLab_NTroje" is the alias for the BMLrub dataset. So please ignore the first question.

Tim-Salzmann avatar Feb 09 '21 11:02 Tim-Salzmann

Thank you for your interests. The forward function of body prior model will first compute the mesh vertex then from those vertex calculate the 3D joints. They are too costly. So that we do not use theirs. But our kinematic tree are the same as theirs.

wei-mao-2019 avatar Mar 12 '21 06:03 wei-mao-2019

hi I should choose that data to download,

AMASS

  1. SMPL+H G
  2. SMPL+x G
  3. SMPL+x N
  4. Render

https://amass.is.tue.mpg.de/download.php

I should choose

whf9527 avatar May 07 '22 06:05 whf9527