search

vchoutas / smplify-x

Expressive Body Capture: 3D Hands, Face, and Body from a Single Image
https://smpl-x.is.tue.mpg.de/
Other
1.75k stars 339 forks source link

Mesh from just openpose keypoints #143

Open nerdygirl21 opened 3 years ago

nerdygirl21 commented 3 years ago

Hi, Is it possible to generate a mesh with just the openpose keypoints information (and not the images folder)?

Wondering if there is a speaker-agnostic way to get these mesh renders. As currently, they rely on the speaker's image and the openpose keypoints.

Looking forward to your reply.

nerdygirl21 commented 3 years ago

@vchoutas Kindly guide me, thank you very much

LuoLuoDev commented 1 year ago

@nerdygirl21 have any answer about generating a mesh with just the openpose keypoints ?

whl-007 commented 2 months ago

Hi, Is it possible to generate a mesh with just the openpose keypoints information (and not the images folder)?

Wondering if there is a speaker-agnostic way to get these mesh renders. As currently, they rely on the speaker's image and the openpose keypoints.

Looking forward to your reply.

你好，是否可以仅使用openpose关键点信息（而不是images文件夹）生成网格？

想 知道是否有一个扬声器不可知的方式来获得这些网格渲染。目前，它们依赖于说话者的图像和openpose关键点。

期待 您的回复。

Can you use own RGB images？I haven't found a way to use SMPLX to turn my own RGB images into SMPLX fits.