DynaAvatar achieves zero-shot 3D human reconstruction from a single image with motion-dependent cloth dynamics.
arXiv · March 17, 2026 · 2603.14772
The Takeaway
Previously, modeling realistic cloth motion required subject-specific optimization or multi-view capture. This feed-forward architecture generalizes to unseen humans and provides realistic physics-aware animations for digital doubles.
From the abstract
Existing single-image 3D human avatar methods primarily rely on rigid joint transformations, limiting their ability to model realistic cloth dynamics. We present DynaAvatar, a zero-shot framework that reconstructs animatable 3D human avatars with motion-dependent cloth dynamics from a single image. Trained on large-scale multi-person motion datasets, DynaAvatar employs a Transformer-based feed-forward architecture that directly predicts dynamic 3D Gaussian deformations without subject-specific o