search

Walter0807 / MotionBERT

[ICCV 2023] PyTorch Implementation of "MotionBERT: A Unified Perspective on Learning Human Motion Representations"
Apache License 2.0
1.02k stars 123 forks source link

Potential error in the application example in readme #56

Closed valentin-fngr closed 1 year ago

valentin-fngr commented 1 year ago

Hi,

this example : https://github.com/Walter0807/MotionBERT#applications states that the input 2d skeleton shoul dbe of shape (N, T, 17, 3). Isn't it (N, T, 17, 2) instead ? Thanks

Walter0807 commented 1 year ago

We use input dim=3 (x, y, confidence).

haofeng0705 commented 9 months ago

@Walter0807 hi author！I want to know if the confidence here is used as an additional dimension to increase estimation accuracy? Will this bring additional computational burden?