MoSE3: Pixel-Level SE(3) Estimation from Monocular Video
MoSE3, a NeurIPS 2026 Spotlight, estimates pixel-level SE(3) motion from monocular video in a single forward pass, adding a tracking branch and differentiable Horn fitting on top of π3.
2026-10-06 ~ 2026-10-08 · 2 related posts
- MoSE3 estimates per-pixel SE(3) transforms from video, NeurIPS Spotlight — du_yilun · 2026-10-06
- MoSE3 recovers per-pixel world-space SE(3) via 3D point tracks and differentiable Horn fit — zhenjun_zhao · 2026-10-08