Get Started
Research questionHow can we predict local 3D occupancy from a single underwater RGB image despite unreliable visual geometry?Underwater appearance often provides weak or misleading depth cues, making it difficult to infer the surrounding 3D free and occupied space from a single image. Errors in this representation can affect navigation and inspection tasks.
Computer Vision
Machine Learning
Robotics
Latest papersRecent research connected to this question, newest first.AquaBEV: Monocular Underwater BEV Occupancy with 3D Sonar SupervisionThe source studies local BEV occupancy prediction from a single underwater RGB image, using paired 3D imaging sonar as training supervision. It reports results on a controlled underwater RGB-to-sonar occupancy benchmark using Visible IoU and Observed IoU.research paper · Sep 3, 2026
Related questions
How can indoor navigation infer a complete metric traversability map from a single egocentric image?How can RGB-D salient-object detection remain reliable when depth measurements are missing or corrupted?How can one 3D semantic occupancy model handle indoor and outdoor scenes with varying cameras, scales, and label taxonomies?How can RGB-D visual pretraining improve 3D awareness without sacrificing semantic transfer?
Home
Topics
Search
Library