Questions about the evaluation strategy of MotionBERT from a reader

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

Questions about the evaluation strategy of MotionBERT from a reader #57

Closed lllluy closed 1 year ago

lllluy commented 1 year ago

Hi，

I have a question about the parameters "2.5d_factor" and "joint_2.5d_image" in the test set. I did not find the relevant information in https://github.com/CHUNYUWANG/lcn-pose#data. Therefore, I wonder how did you get it? Why don't the output of the model be evaluated directly with the "joint3d_images" calculation MPJPE?

Thanks in advance.

Walter0807 commented 1 year ago

Hi, please refer to https://github.com/Walter0807/MotionBERT/issues/51 which could help.