Depth3DLane combines a depth-deconstruction head, depth-prior distillation from Depth Anything V2, and a CRF refiner to reach reported state-of-the-art z-axis accuracy in monocular 3D lane detection.
A multi-task vision transformer for segmentation and monocular depth estimation for autonomous vehicles
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
-
Depth3DLane: Monocular 3D Lane Detection via Depth Prior Distillation
Depth3DLane combines a depth-deconstruction head, depth-prior distillation from Depth Anything V2, and a CRF refiner to reach reported state-of-the-art z-axis accuracy in monocular 3D lane detection.