.png){kind=icon}
.png){kind=icon}
Return
SharpEdge: High-quality data-driven monocular depth estimation for enhanced boundary precision
.png){kind=icon}
PRE
AI
DOI：10.1016/j.engappai.2025.112067.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
While existing monocular depth estimation methods have achieved commendable performance, they often fall short in accurately distinguishing object boundaries. This deficiency largely stems from the inherent noise in dataset acquisition, such as unclear edges and missing depth information. To address these challenges, this paper introduces a novel, high-quality, data-driven monocular depth estimation method tailored for autonomous driving. The approach significantly enhances depth predictions with clearer object boundaries and reduced noise, making it well-suited for real-time, safety-critical applications.
Keywords:
monocular depth estimation
object boundaries
autonomous driving
data-driven method
depth prediction
Journal
IF:
8
Papers:
5.3K
Citations:
3.5W