Sparkle: A Robust and Versatile Representation for Point Cloud based Human Motion Capture
This addresses the problem of balancing expressiveness and robustness in point cloud-based motion capture for applications like privacy-preserving sensing, though it appears incremental as it builds on existing point-based and skeleton-based methods.
The paper tackles the challenge of learning robust representations from noisy point clouds for human motion capture by proposing Sparkle, a structured representation that unifies skeletal joints and surface anchors with kinematic-geometric factorization. The result is state-of-the-art performance in accuracy, robustness, and generalization under domain shifts, noise, and occlusion.
Point cloud-based motion capture leverages rich spatial geometry and privacy-preserving sensing, but learning robust representations from noisy, unstructured point clouds remains challenging. Existing approaches face a struggle trade-off between point-based methods (geometrically detailed but noisy) and skeleton-based ones (robust but oversimplified). We address the fundamental challenge: how to construct an effective representation for human motion capture that can balance expressiveness and robustness. In this paper, we propose Sparkle, a structured representation unifying skeletal joints and surface anchors with explicit kinematic-geometric factorization. Our framework, SparkleMotion, learns this representation through hierarchical modules embedding geometric continuity and kinematic constraints. By explicitly disentangling internal kinematic structure from external surface geometry, SparkleMotion achieves state-of-the-art performance not only in accuracy but crucially in robustness and generalization under severe domain shifts, noise, and occlusion. Extensive experiments demonstrate our superiority across diverse sensor types and challenging real-world scenarios.