search

HongwenZhang / PyMAF

[ICCV 2021, Oral] PyMAF: 3D Human Pose and Shape Regression with Pyramidal Mesh Alignment Feedback Loop
https://hongwenzhang.github.io/pymaf
Other
594 stars 67 forks source link

human36m GT SMPL parameter #10

Closed sunwonlikeyou closed 3 years ago

sunwonlikeyou commented 3 years ago

You said "The ground truth SMPL parameters in Human3.6M are generated by applying MoSH [34] to the sparse 3D MoCap marker data, as done in Kanazawa et al. " I have angle files,3D points of Human3.6M. But i dont know how to SMPL parameters from the dataset. Do you have code ? or Is there other data from Human3.6M??

HongwenZhang commented 3 years ago

Hi, we used the moshed labels shared in HMR. Unfortunately, the moshed labels can not be redistributed publicly due to the license issue. However, this repo may still help you get the data and labels.

sunwonlikeyou commented 3 years ago

Thank you for reply! Shared link just give me 3D,2D poses, but I want to get GT SMPL parameters of Human36M. So.. Can I receive moshed labels of Human36M..? It's just research purpose and I want to compare it with pseudo gt parameters.

Thank you very much. Have a nice day.

2021년 6월 16일 (수) 오후 9:31, Hongwen Zhang @.***>님이 작성:

Hi, we used the moshed labels shared in HMR https://github.com/akanazawa/hmr/blob/master/doc/train.md#download-datasets. Unfortunately, the moshed labels can not be redistributed publicly due to the license issue. However, this repo https://github.com/mks0601/3DMPPE_ROOTNET_RELEASE may still help you get the data and labels.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/HongwenZhang/PyMAF/issues/10#issuecomment-862339350, or unsubscribe https://github.com/notifications/unsubscribe-auth/AM4NTKF4ZPCJQ74NIHXMLX3TTCKQXANCNFSM46Y752TA .