Get Started
Research questionHow can inertial measurements recover metric scale in pretrained monocular 3D predictions?Monocular images do not reveal the real-world size of a scene, so pretrained 3D models can estimate geometry and camera motion while misjudging absolute scale. Inertial measurements provide motion information in metric units that may resolve this ambiguity.
Computer Vision
Machine Learning
Latest papersRecent research connected to this question, newest first.VI3: Grounding Pretrained 3D Foundation Models with Inertial CuesApplies to pretrained monocular 3D foundation models that predict camera poses and dense depth, with IMU readings available. The evidence covers synthetic and real aerial datasets and reports scale recovery while preserving geometric consistency; inertial motion is more informative in some conditions than others.research paper · Sep 3, 2026
Related questions
How can single-image depth estimation resolve scale ambiguity to produce consistent metric depth across environments?How can indoor navigation infer a complete metric traversability map from a single egocentric image?How can we accurately track an athlete’s 3D center of mass from a single phone camera?How can neural map matchers learn accurate 3-DoF poses from noisy GPS positions and heading labels?
Home
Topics
Search
Library