Map-Det3D: Metric Feed-Forward 3D Reconstruction Prior for Multi-view 3D Object Detection
rsasaki0109 · x · 2026-08-18
Map-Det3D is a metric feed-forward 3D reconstruction prior for multi-view 3D object detection from streaming RGB inputs, accepted to ECCV 2026. The method maps a short temporal window into multiple views and repurposes a feed-forward metric 3D reconstruction model as its geometric backbone, predicting boxes directly in metric 3D space without the widely used 2D-to-3D lifting. Experiments show strong online performance and robust transfer without adaptation.
More from Embodied
- MolmoMotion forecasts 3D point trajectories via language instructions — allen_ai · 2026-08-18
- University of Michigan open-sources entire robotics curriculum — lukas_m_ziegler · 2026-08-18
- Meituan demonstrates commercial drone food delivery in Chinese cities — rohanpaul_ai · 2026-08-18
- NVIDIA Thor edge PC runs RealSense D457 with local Ollama VLM — chrismatthieu · 2026-08-18
- MIT Research: Autonomous Lab + Generative Model for Enzyme Engineering — ProfBuehlerMIT · 2026-08-18
- Investment thesis on the robotics market: inflection point and scale — Rewkang · 2026-08-18