search

Daniil-Osokin / lightweight-human-pose-estimation-3d-demo.pytorch

Real-time 3D multi-person pose estimation demo in PyTorch. OpenVINO backend can be used for fast inference on CPU.
Apache License 2.0
653 stars 137 forks source link

Does parse_poses poses_2d always return the full array of joints? #93

Closed adammpolak closed 2 years ago

adammpolak commented 2 years ago

Does poses_2d return an array of all joints and their 2D locations or only the joints that are detected/not occluded in the image?

https://github.com/Daniil-Osokin/lightweight-human-pose-estimation-3d-demo.pytorch/blob/8794edc873d8db4f582dcc8a578f99bf4e218830/demo.py#L99

Daniil-Osokin commented 2 years ago

Yes, you can check the code.

yanglilongss commented 2 years ago

Hello, the prediction effect of the provided model human body side angle is not very good. Is there a better prediction model

Daniil-Osokin commented 2 years ago

@yanglilongss Please, check #94.

Daniil-Osokin commented 2 years ago

Hope, it helped.