Neural Human Pose Prior
This work provides a probabilistic foundation for integrating pose priors into human motion capture and reconstruction pipelines, addressing a domain-specific need in computer vision and graphics.
The paper tackled the problem of modeling a neural prior over human body poses by introducing a data-driven approach using normalizing flows, specifically RealNVP, to learn a flexible density over poses in the 6D rotation format, with results demonstrated through qualitative and quantitative evaluations.
We introduce a principled, data-driven approach for modeling a neural prior over human body poses using normalizing flows. Unlike heuristic or low-expressivity alternatives, our method leverages RealNVP to learn a flexible density over poses represented in the 6D rotation format. We address the challenge of modeling distributions on the manifold of valid 6D rotations by inverting the Gram-Schmidt process during training, enabling stable learning while preserving downstream compatibility with rotation-based frameworks. Our architecture and training pipeline are framework-agnostic and easily reproducible. We demonstrate the effectiveness of the learned prior through both qualitative and quantitative evaluations, and we analyze its impact via ablation studies. This work provides a sound probabilistic foundation for integrating pose priors into human motion capture and reconstruction pipelines.