MVTracker is a feed-forward multi-view 3D point tracker that fuses depth-lifted features into a point cloud with kNN correlation and a transformer, trained on synthetic data and evaluated on real benchmarks with errors of a few centimeters.
Easi3R: Estimating disentangled motion from dust3r without training
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.CV 1years
2025 1verdicts
CONDITIONAL 1roles
background 1polarities
unclear 1representative citing papers
citing papers explorer
-
Multi-View 3D Point Tracking
MVTracker is a feed-forward multi-view 3D point tracker that fuses depth-lifted features into a point cloud with kNN correlation and a transformer, trained on synthetic data and evaluated on real benchmarks with errors of a few centimeters.