GVHMR: World-Grounded Human Motion Recovery via Gravity-View Coordinates
rsasaki0109 · x · 2026-08-26
GVHMR (World-Grounded Human Motion Recovery via Gravity-View Coordinates) is a paper accepted at Siggraph Asia 2024 and TPAMI 2026.
Core Problem: Recovering world-grounded global human motion from monocular videos, specifically addressing error accumulation in long videos.
Key Idea: Introducing a novel "Gravity-View Coordinates" system, inferred naturally from gravity and camera-view directions in each image, to improve pose estimation accuracy.
Resources: The model is trained on AMASS, BEDLAM, H36M, and 3DPW datasets. All codes and weights are publicly available.
Related event: GVHMR: Open-Source Recovery of Global Human Motion from Monocular Video(2 posts)→
More from Research
- Scaling Laws Extrapolation Unreliable Due to Statistical Reasons — DanielKhashabi · 2026-08-26
- KostasVisualizations: Interactive Visualizations for Computer Vision and Machine Learning — CSProfKGD · 2026-08-26
- AI Scholar Praises Model Architecture Research: Value of 180B Scale Validation and Ablation Studies — kastnerkyle · 2026-08-26
- FixAnything: Unified 3D Rendering Artifact Repair via Video Generative Priors — gabriel1 · 2026-08-26
- Paper proposes a fix for running context recirculation issues — yogthos · 2026-08-26
- Study Finds Attention-Mask Checks Miss Causality Leaks in Sequence Models — VIDraft · 2026-08-26