zc-alexfan/arctic

Hand-object mesh mismatch with the manopth

GuanHY opened this issue · 2 comments

GuanHY commented

Hi Alex,

I just tried to use the differentiable manolayer here: https://github.com/hassony2/manopth/blob/master/manopth/manolayer.py, to reproduce the hand-object mesh with the provided beta and theta parameters in arctic_data/data/raw_seqs(https://github.com/zc-alexfan/arctic/blob/master/docs/data/README.md). But I find that the visualization results are always abnormal, such as fingers.
7d0f5d2eeb63c8ab956cc64e91d554c
ff6f31479eaac3bb2916459016a60bc
54daf3035c5e5ba3ff69006693e204f
e906e293c0cdf9d85df5bf7846d5550

I've noticed that you use AITviewer for visualization. I'd like to ask if there are any notable differences worth paying attention to?Do you have any guess on the cause for this, or have you tried to use their manolayer before?

zc-alexfan commented

Sorry, we mainly tested out with smplx mano model. There was an issue raised a year ago which seems to show similar issues but I have not used the mano layer from hasson before.

GuanHY commented

Yes, I've also read that issue and it appears that it was not resolved in the end.
Thank you very much for your response. I will look into the SMPLX mano model to see if it can help address the issue we're facing