how to lift the 2D joint positions to 3D?

[ZJYCP]

hi, I have extracted the 2D positions using OpenPose, but I have no idea about lifting it to 3D as you mentioned in the paper. Could you please provide the code or give me some tips? Besides, I wonder how does the "trg" file arrange 150 positions data.
thanks :-)

[BenSaunders27]

Hi ZJYCP,

The 3D positions are created using the code provided at https://github.com/gopeith/SignLanguageProcessing under 3DposeEstimator.

The "trg" file contains skeleton data of each frame, with a space separating frames. Each frame contains 150 joint values and a subsequent counter value, all separated by a space. If your data contains 150 joints per frame, please ensure that trg_size is set to 150 in the config file.

[ZJYCP]

thanks for your reply.
I have used '3DposeEstimator' to create 3D positions. But the results are a little different from my expectation, which are showed below.

lots of number are larger than 1, which is quiet different with the tmp data you provide. And in this way, the generated video also makes no sense.

Here is what I did:

1. using openpose python api to estimate the keypoints. Here I didn't set param 'keypoint_scale' in openpose.
2. then I followed the pipeline-demo in https://github.com/gopeith/SignLanguageProcessing , firstly openpose data to h5. and then run pipeline_demo_02_filtr.py, which is same with the demo.py file in 3DposeEstimator . And I got the result as showed above.

Did I miss something or done something wrong? Data processing really takes me a lot of time, ( •́ω•̩̥̀ ) waiting for you reply.
thanks:-)

[BenSaunders27]

Hi,

When I originally processed the data, I divided the joints outputs of the Inverse Kinematics by 3, in order to get them all below 1. Looking at your example, it seems the hands are correct but large, which is explained by the larger scale you use.

Apologies, this was an error on my part, I'll update the README file to explain this.

[hacker009-sudo]

@ZJYCP , please share the pretrained model, if you have stored it somewhere. Request you to please help it is really really urgent.