URNet: Uncertainty-aware Refinement Network for Event-based Stereo Depth Estimation

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Cheng, Yifeng, Knoll, Alois, Cao, Hu
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866911169196851200
author Cheng, Yifeng
Knoll, Alois
Cao, Hu
author_facet Cheng, Yifeng
Knoll, Alois
Cao, Hu
contents Event cameras provide high temporal resolution, high dynamic range, and low latency, offering significant advantages over conventional frame-based cameras. In this work, we introduce an uncertainty-aware refinement network called URNet for event-based stereo depth estimation. Our approach features a local-global refinement module that effectively captures fine-grained local details and long-range global context. Additionally, we introduce a Kullback-Leibler (KL) divergence-based uncertainty modeling method to enhance prediction reliability. Extensive experiments on the DSEC dataset demonstrate that URNet consistently outperforms state-of-the-art (SOTA) methods in both qualitative and quantitative evaluations.
format Preprint
id arxiv_https___arxiv_org_abs_2509_18184
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle URNet: Uncertainty-aware Refinement Network for Event-based Stereo Depth Estimation
Cheng, Yifeng
Knoll, Alois
Cao, Hu
Computer Vision and Pattern Recognition
Event cameras provide high temporal resolution, high dynamic range, and low latency, offering significant advantages over conventional frame-based cameras. In this work, we introduce an uncertainty-aware refinement network called URNet for event-based stereo depth estimation. Our approach features a local-global refinement module that effectively captures fine-grained local details and long-range global context. Additionally, we introduce a Kullback-Leibler (KL) divergence-based uncertainty modeling method to enhance prediction reliability. Extensive experiments on the DSEC dataset demonstrate that URNet consistently outperforms state-of-the-art (SOTA) methods in both qualitative and quantitative evaluations.
title URNet: Uncertainty-aware Refinement Network for Event-based Stereo Depth Estimation
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2509.18184