Closed Xuehao-Gao closed 1 year ago
As what is argured by Wei Mao et al., angle representation is ambiguous, you could find it in the Fig 4 of their paper, that is the reason why we use 3D coordinate representation, instead of the angle representation .
@inproceedings{mao2019learning, title={Learning trajectory dependencies for human motion prediction}, author={Mao, Wei and Liu, Miaomiao and Salzmann, Mathieu and Li, Hongdong}, booktitle={Proceedings of the IEEE/CVF International Conference on Computer Vision}, pages={9489--9497}, year={2019} }
Can you provide the official MAE performance of MSRGCN to help others compare it with other methods more comprehensively?
Many Thanks