.png){kind=icon}
.png){kind=icon}
Return
ConvDepth: Self-supervised monocular depth estimation for autonomous driving
.png){kind=icon}
PRE
AI
DOI：10.1016/j.neucom.2026.133681.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
• Propose ConvDepth, a CNN-based self-supervised MDE framework. • Introduce multi-level detail enhancement for fine-grained features. • Develop a probabilistic disparity head to mitigate depth regression ambiguities. • Formulate a robust self-distillation loss to filter noise in pseudo-labels.
Keywords:
ConvDepth
self-supervised monocular depth estimation
multi-level detail enhancement
probabilistic disparity head
self-distillation loss
Journal
IF:
6.5
Papers:
2.5W
Citations:
6.5W