muelea / selfcontact

CVPR 2021 - Official code repository for the paper: On Self-Contact and Human Pose.
https://tuch.is.tue.mpg.de
Other
86 stars 6 forks source link

running run_selfcontact_optimization.py in SMPL model #9

Open qinb opened 2 years ago

qinb commented 2 years ago

Hi, I successfully run the run_selfcontact_optimization.py in model_type='smplx', when I change model_type from smplx to smpl, and modify some codes to be compatible with it, the mismatch still occurs. Finally, It may be a solution to utilize the smpl_to_smplx.pkl to convert to smplx by smpl model, but I don't how to use it. Could you offer some help? Thanks @paulchhuang @ahmedosman

muelea commented 2 years ago

Hi qinb,

thanks for your questions! I never tested the selfcontact-optimization for SMPL meshes. Even if the code runs, you might need to adjust some weights later on if you wanted to use SMPL. If you would like to fit SMPL-X to you SMPL meshes, you can follow the instructions in this repo: https://github.com/vchoutas/smplx/tree/master/transfer_model

Best wishes, Lea

qinb commented 2 years ago

Thanks, I have utilized this repo to exchange from smpl to smplx. The result looks good but the speed is too slow. I have several questions:

  1. How to accelerate, Except decrease the parameter - maxiter, what has the other methods?
  2. ProHMR serves as a pose estimator, to solve self contact question, does your method merge together with ProHMR for joint training? for example, your method serves as a loss supervisor? Thanks, again @muelea @paulchhuang