search

Vegetebird / MHFormer

[CVPR 2022] MHFormer: Multi-Hypothesis Transformer for 3D Human Pose Estimation
MIT License
530 stars 85 forks source link

About 2D attitude detection. #97

Closed Ared521 closed 1 year ago

Ared521 commented 1 year ago

Thank you for your great work. I have a question to ask. I see that the 2D pose detection in your code uses HRNet, which can only detect a single person. If I want to detect the 3D pose of multiple people, can I use HigherHRNet instead? , detect the 2D poses of multiple people, and then use MHFormer to detect the 3D poses of them?

Vegetebird commented 1 year ago

Yes, you can use HigherHRNet. Alternatively, you can detect the bounding boxes of multiple people using YOLOv3, track each person, and then feed the tracked bounding boxes of each person into HRNet.

Ared521 commented 1 year ago

Yes, you can use HigherHRNet. Alternatively, you can detect the bounding boxes of multiple people using YOLOv3, track each person, and then feed the tracked bounding boxes of each person into HRNet.

OK, Thank you!