Related papers appearing in ICCV'23/NeurIPS'23; Recent papers missed out in the survey. about hmr-survey HOT 1 OPEN

The following papers have been updated to the arXiv version of the survey [v6, Jan.2, 2023].

Datasets
SynBody: Synthetic Dataset with Layered Human Models for 3D Human Perception and Modeling
Motion-X: A Large-scale 3D Expressive Whole-body Human Motion Dataset

Whole-Body Mesh Recovery
SMPLer-X: Scaling Up Expressive Human Pose and Shape Estimation
Towards Robust and Expressive Whole-body Human Pose and Shape Estimation

Diffusion for HMR
Distribution-Aligned Diffusion for Human Mesh Recovery
Probabilistic Human Mesh Recovery in 3D Scenes from Egocentric Views
Generative Approach for Probabilistic Human Mesh Recovery using Diffusion Models

Transformer for HMR
TORE: Token Reduction for Efficient Human Mesh Recovery with Transformer
JOTR: 3D Joint Contrastive Learning with Transformers for Occluded Human Mesh Recovery

Domain Adaptation
Cyclic Test-Time Adaptation on Monocular Video for 3D Human Mesh Reconstruction

Output Type
3D Human Mesh Recovery with Sequentially Global Rotation Estimation

Camera Setting
Zolly: Zoom Focal Length Correctly for Perspective-Distorted Human Mesh Reconstruction

from hmr-survey.