A transformer-based variational motion prior trained on large-scale 3D human motion data can be frozen and reused, with lightweight adapters, for pose estimation from depth, LiDAR, and IMU data and for filling missing frames.
Neural marionette: Unsupervised learning of motion skeleton and latent dynamics from vol- umetric video
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2024 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
ReMP: Reusable Motion Prior for Multi-domain 3D Human Pose Estimation and Motion Inbetweening
A transformer-based variational motion prior trained on large-scale 3D human motion data can be frozen and reused, with lightweight adapters, for pose estimation from depth, LiDAR, and IMU data and for filling missing frames.