F²SLAM Turns Feed-Forward Geometry into Persistent Factors, Cutting Replica ATE to 0.002 m
zhenjun_zhao · x · 2026-10-19
A new paper (arXiv:2610.05207) by Zhisong Xu, Fan Zhu et al., with Javier Civera, presents F²SLAM: instead of treating feed-forward 3D predictions as an external state fused after the fact, it converts them into optimization-native target-weight measurements attached to a persistent dense factor graph.
- A high-frequency stream maintains local tracking and graph connectivity; a low-frequency stream uses wider multi-view context to selectively refresh measurements after a state-consistency check
- Both streams constrain poses, inverse depths and optional intrinsics via a single dense bundle adjustment
- In the uncalibrated setting, average ATE RMSE on Replica drops from 0.030 m (strongest feed-forward baseline) to 0.002 m
- Consistent gains in trajectory estimation and dense reconstruction across benchmarks; code to be released upon acceptance
More from Embodied
- F-22 controls a CCA drone for the first time, Anduril and USAF hit milestone — BrettKrieger12 · 2026-10-20
- Ropedia W0 shows 10k hours of human video yields a robot learning scaling law — liuziwei7 · 2026-10-20
- Apple Sets October 28 Event, Touchscreen MacBook Pro Reportedly Unveiled — Polymarket · 2026-10-20
- Weekend DIY: two RP2040 boards give AI agents keyboard-video-mouse control of phones — hugs · 2026-10-20
- LidarVerse: one real-time LiDAR world model for reconstruction, forecasting and generation — ZGojcic · 2026-10-20
- AI Lab Offers Free Plumbing and HVAC Repairs to Film Wearable-Sensor Video for Robot Training — whatdowecallu · 2026-10-20