REMOTE combines optical flow, scene features, and attention-based joint features to predict endoscopic camera motion, reporting state-of-the-art accuracy on three datasets at real-time speed.
Location recognition using prioritized feature matching,
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
-
REMOTE: Real-time Ego-motion Tracking for Various Endoscopes via Multimodal Visual Feature Learning
REMOTE combines optical flow, scene features, and attention-based joint features to predict endoscopic camera motion, reporting state-of-the-art accuracy on three datasets at real-time speed.