CVSep 18, 2025

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

arXiv:2509.18184v12 citationsh-index: 7Visual Intelligence
Originality Incremental advance
AI Analysis

This work addresses depth estimation for event cameras, which is important for robotics and autonomous systems, but appears incremental as it builds on existing event-based stereo methods.

The paper tackles event-based stereo depth estimation by introducing URNet, an uncertainty-aware refinement network that incorporates local-global refinement and KL divergence-based uncertainty modeling, achieving state-of-the-art performance on the DSEC dataset.

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.

Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes