An efficient transformer architecture for BEV instance prediction reduces parameter counts and inference times versus SOTA by relying on a simplified paradigm of only instance segmentation and flow prediction.
Beverse: Unified perception and prediction in birds-eye-view for vision-centric autonomous driving
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2024 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Fast and Efficient Transformer-based Method for Bird's Eye View Instance Prediction
An efficient transformer architecture for BEV instance prediction reduces parameter counts and inference times versus SOTA by relying on a simplified paradigm of only instance segmentation and flow prediction.