AtomicMotion partitions the body into five clusters, applies masked full-body pre-conditioning, and introduces Kinematic Attention to reconstruct poses from head and hand data, claiming better fidelity than baselines on AMASS.
Realistic full-body tracking from sparse observations via joint-level modeling
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
AtomicMotion: Learning Human Motion From Different Human Parts
AtomicMotion partitions the body into five clusters, applies masked full-body pre-conditioning, and introduces Kinematic Attention to reconstruct poses from head and hand data, claiming better fidelity than baselines on AMASS.