search

SangbumChoi / MobileHumanPose

This repo is official PyTorch implementation of MobileHumanPose: Toward real-time 3D human pose estimation in mobile devices(CVPRW 2021).
MIT License
232 stars 29 forks source link

Why do I need to load 2D data when training 3D data？ #2

Closed APeiZou closed 3 years ago

APeiZou commented 3 years ago

image Why do I need to load 2D MPII dataset when training the 3D Human36M dataset？

SangbumChoi commented 3 years ago

@APeiZou Theoritically you don't need to add 2D dataset while training 3D dataset. However, it is widely known that additional 2D dataset without z axis is giving more robustness to the 3D human pose estimation model due to seeing diverse dataset. You can see more detail at maybe https://arxiv.org/abs/1711.08229 (hope i linked appropriate paper).