Closed-Form Rescaling Fixes World-Space Human Drift in Fast-Motion Dance Videos
Hao-Jen Chien · hf · 2026-10-16
Researchers from National Yang Ming Chiao Tung University and Boston University publish a diagnostic study on world-space human recovery under high camera dynamics.
Key findings:
- In fast-motion dance videos, independently recovered camera and human estimates suffer a scale mismatch, causing severe world-space drift/sliding in multi-person reconstruction
- The authors propose a closed-form rescaling method that corrects the mismatch without retraining
- Demonstrated on K-Pop broadcast footage and other high-dynamics multi-person scenes; project page worldhmr-fastcam is public
The contribution is attributing reconstruction drift to a diagnosable scale factor and offering a plug-and-play, training-free fix.
More from Research
- Embedding arithmetic demo: cartoon − animation − comic = diagram — mayfer · 2026-10-23
- Omnii Genome: Single Generalist Model Outperforms AlphaGenome Across All Benchmarks — JosephJacks_ · 2026-10-23
- Domingos Cites New Result Claiming RL Is 'Basically Useless' vs Supervised Learning — pmddomingos · 2026-10-23
- Word Arithmetic on Embeddings: "Bathroom Minus Toilet Equals Kitchen" Beats SAE Probing — mayfer · 2026-10-23
- Wolfram Opens Rule 30 Prizes to AIs: $30,000 for Three Open Problems — cephaloform · 2026-10-23
- Stanford and Databricks use AI agents to reproduce medical findings and test analytical robustness — ChengleiSi · 2026-10-23