Search papers, labs, and topics across Lattice.
AtomicMotion addresses the problem of reconstructing full-body human poses from sparse head and hand trajectories by decoupling and reintegrating body dynamics. The method partitions the skeleton into five clusters based on functional intent, uses a masked full-body pre-conditioning strategy during training, and introduces Kinematic Attention to ensure biological plausibility. Experiments on the AMASS dataset show that AtomicMotion outperforms existing baselines in reconstruction fidelity and biomechanical realism.
By explicitly modeling the kinematic structure of the human body, AtomicMotion achieves more realistic and accurate full-body motion reconstruction from sparse inputs than monolithic approaches.
Accurately reconstructing full-body poses from sparse head and hand trajectories is a foundational challenge for immersive AR/VR telepresence. Current methods often struggle with error accumulation and unnatural joint coordination, primarily because they treat the human body as a monolithic entity, thereby failing to capture the fine-grained ``atomic intents'' embedded in subtle signal variations and overlooking the inherent structural topology. To bridge this gap, we present AtomicMotion, a framework designed to decouple and re-integrate body dynamics through three core innovations. First, we introduce a logical body partitioning scheme that decomposes the skeleton into five distinct clusters based on functional intent; this ensures that each partition preserves internal joint synergies while isolating local motion primitives. Second, to robustly map sparse inputs to high-dimensional poses, we employ a masked full-body pre-conditioning strategy during training, forcing the model to internalize global skeletal topology and latent kinematic constraints. Finally, addressing the limitations of vanilla spatial attention, which often ignores fixed physiological connectivity, we propose Kinematic Attention. By embedding the classical kinematic tree structure into the attention mechanism, we ensure biological plausibility in the synthesized motions. Extensive evaluations on the AMASS dataset demonstrate that AtomicMotion significantly outperforms existing baselines, yielding higher reconstruction fidelity and superior biomechanical realism.
