Open wpf89928 opened 1 year ago
Sorry for the late reply. The hand models currently used in MMPose employ top-down methods, which consider the area within the hand bounding boxes as a single hand, even if there are two hands present in the box. You can try RTMW models for wholebody pose estimation, handle scenarios with two hands
What is the feature?
Hi, very excellent work!
Just curious if any solution for the case where the two hands are overlapping or crossing.
I tried the published hand-5 model, and it only gives keypoints of one hand. Can you advise how to solve this case?
Thanks!
Any other context?
No response