Reproduce “DUNEDepth: Bronchoscopic Depth and Pose Estimation via Hybrid Adaptation of a 3D-Vision Foundation Model” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Qi Chang, Austin Kao, Ange Lou, William E. Higgins
DUNEDepth estimates airway depth and camera motion from neighboring bronchoscopy frames, combining a 3D-vision model with CT-derived virtual-view depth supervision and video-derived motion cues. On 42 viewpoints, it has lower L1/L2 depth errors than SC-SfMLearner, but not the best median-scaled L1. On sequence C, trajectory error is 3.84 mm versus 8.61 mm; registration uncertainty is below 3 mm.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.