DOPE augments a DUET navigation backbone with text and image object attention modules and reports improved success rates on the R2R and REVERIE benchmarks.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
DOPE: Dual Object Perception-Enhancement Network for Vision-and-Language Navigation
DOPE augments a DUET navigation backbone with text and image object attention modules and reports improved success rates on the R2R and REVERIE benchmarks.