The submission adapts OmniDrive with a DVPE-style divided-view perception module to enhance instruction-conditioned ego trajectory prediction on nuScenes scenes for the doScenes challenge.
Dc-sam: In-context segment anything in images and videos via dual consistency
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
A DVDrive Approach for doScenes Instructed Driving Challenge
The submission adapts OmniDrive with a DVPE-style divided-view perception module to enhance instruction-conditioned ego trajectory prediction on nuScenes scenes for the doScenes challenge.