A fixed-size triplane scene representation tokenizes any number of camera images into 45-104 tokens, enabling faster autoregressive driving policies without sacrificing planning accuracy.
Barron, Ben Mildenhall, Dor Verbin, Pratul P
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
-
Efficient Multi-Camera Tokenization with Triplanes for End-to-End Driving
A fixed-size triplane scene representation tokenizes any number of camera images into 45-104 tokens, enabling faster autoregressive driving policies without sacrificing planning accuracy.