F²SLAM Benchmarks: 0.002 m ATE on Uncalibrated Replica at 7.4 FPS on One A100
zhenjun_zhao · x · 2026-10-19
The authors share detailed benchmark results for F²SLAM across TUM RGB-D, 7-Scenes, Replica and KITTI:
- Best average ATE in the uncalibrated setting: 0.027 m on TUM RGB-D and 0.002 m on Replica
- Well ahead of uncalibrated foundation-model baselines on KITTI
- 7-Scenes reconstruction vs SLAM-Former: completion 0.030→0.015 m, Chamfer 0.027→0.019 m
- 7.4 FPS on a single A100 with lower peak GPU memory (15.4 vs 27.7 GB)
Architecture: a two-clock factor graph where a high-frequency stream (DROID-Net) maintains local tracking and connectivity, while a low-frequency stream (DINO + Flow3r) predicts displacement and confidence over a wider window and refreshes existing edges, gated by a state-consistency check that locks replaced factors.
More from Embodied
- Robots ace blind precision assembly using only joint encoder signals, no cameras — kevin_zakka · 2026-10-19
- If AI records every conversation by default, society changes dramatically — generativist · 2026-10-19
- Tesla shareholder letter teases a mysterious Optimus v3 silhouette — CyberRobooo · 2026-10-19
- Robotics is having its GPT-2 moment, says State of AI's Nathan Benaich — nathanbenaich · 2026-10-19
- Robotics founder's China day 6: not the top AI innovator, but adoption will lead by a massive gap — DominiqueCAPaul · 2026-10-19
- Ropedia Kicks Off Super Week With W0, a World Model Teaching Robots to Predict Physics Before Acting — liuziwei7 · 2026-10-19