Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Keypoints Conversion - SMPL to COCO #70

Open
Dipankar1997161 opened this issue Jul 30, 2023 · 1 comment
Open

Keypoints Conversion - SMPL to COCO #70

Dipankar1997161 opened this issue Jul 30, 2023 · 1 comment

Comments

@Dipankar1997161
Copy link

Hello @hongsukchoi,

Hope you are doing well.

I generated the Smpl and successfully run the 3d reconstruction on those points.

My final task is - extracting the smpl key points (which are the original smpl joints) and then performing 2d and 3d pose estimation on the data.

Q1. The smpl key points are the "Joints" or the "Poses" parameter in Smpl?
Q2. How do I extract the smpl key points and convert them to coco format?

Can you provide some insights on this? Any repo or Pose2Mesh can provide any such thing?

I am okay with both methods but haven't found many reliable sources for the conversion of SMPL to COCO.

I will wait for your response.

@hongsukchoi
Copy link
Owner

Hi @Dipankar1997161, sorry for the late reply.

Q1. The smpl 'pose' parameters are joint angles (not joint position) in axis-angle representation. The smpl key points you are referring to are probably the "Joints" (joint position).
Q2. You could estimate SMPL pose paramters from the smpl keypoints and reconstruct the 3D mesh, and multiply the coco joint matrix to the mesh vertices to get the coco joints.
coco joint matrix: https://github.com/hongsukchoi/Pose2Mesh_RELEASE/blob/master/data/COCO/J_regressor_coco.npy
This joint matrix is approximately performing forward kinematics.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants