search

open-mmlab / mmpose

OpenMMLab Pose Estimation Toolbox and Benchmark.
https://mmpose.readthedocs.io/en/latest/
Apache License 2.0
5.89k stars 1.26k forks source link

How to transform 3D keypoints from a coordinate system centered on the pelvis to a camera coordinate system #2794

Open hirokiyazaki opened 1 year ago

hirokiyazaki commented 1 year ago

📚 The doc issue

Hello, I have run the code example provided below for the "3D Human Pose Demo" and acquired the 3D keypoints coordinates for each frame.

python demo/body3d_pose_lifter_demo.py \ demo/mmdetection_cfg/rtmdet_m_640-8xb32_coco-person.py \ https://download.openmmlab.com/mmpose/v1/projects/rtmpose/rtmdet_m_8xb32-100e_coco-obj365-person-235e8209.pth \ configs/body_2d_keypoint/rtmpose/body8/rtmpose-m_8xb256-420e_body8-256x192.py \ https://download.openmmlab.com/mmpose/v1/projects/rtmposev1/rtmpose-m_simcc-body7_pt-body7_420e-256x192-e48f03d0_20230504.pth \ configs/body_3d_keypoint/video_pose_lift/h36m/video-pose-lift_tcn-243frm-supv-cpn-ft_8xb128-200e_h36m.py \ https://download.openmmlab.com/mmpose/body3d/videopose/videopose_h36m_243frames_fullconv_supervised_cpn_ft-88f5abbb_20210527.pth \ --input https://user-images.githubusercontent.com/87690686/164970135-b14e424c-765a-4180-9bc8-fa8d6abc5510.mp4 \ --output-root vis_results \ --save-predictions

However, since these coordinates are within a coordinate system centered around the pelvis, I would like to transform the keypoints' coordinates into the camera coordinate system. Therefore, could you please provide me with guidance on how to perform this transformation?

Suggest a potential alternative/fix

No response

LZhen624 commented 1 year ago

您好，请问您的3D 人体姿态估计能运行出来吗，我搞了好几天这个都弄不了，能否相互交流一下呢

LZhen624 commented 1 year ago

Hello, may I ask if your 3D human posture can be estimated to work? I've been working on this for several days but I can't figure it out. Can you please communicate with each other

hirokiyazaki commented 1 year ago

Sure, no problem!

LZhen624 commented 1 year ago

Do you know which ready-made model can also mark several key points of the foot？I want to demonstrate the key points of the whole body and feet in the 3D human pose. There are not many key points of the feet, just have them.

hirokiyazaki commented 1 year ago

I'm sorry, I don’t know such an existing model. I apologize for not being able to help.

LZhen624 commented 1 year ago

Thank you very much. Recently, I found the Halpe26 model in the RTM pose in mmpose, which includes key points of the feet. Can you use it,