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

2025/09/18 by Yifeng Cheng, Alois Knoll, Cheng, Yifeng +3 · 1 citation
Engineering · Computer Science · #Advanced Memory and Neural Computing #Advanced Vision and Imaging #Generative Adversarial Networks and Image Synthesis

paper · pdf · doi:10.48550/arxiv.2509.18184

Abstract

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.

Citations

• Stereo Anywhere: Robust Zero-Shot Deep Stereo Matching Even Where Either Stereo or Mono Fail
• LiDAR-Event Stereo Fusion with Hallucinations
• Embracing Events and Frames with Hierarchical Feature Refinement Network for Object Detection
• Temporal Event Stereo via Joint Learning with Stereoscopic Flow
• Demystify Mamba in Vision: A Linear Attention Perspective
• Playing to Vision Foundation Model's Strengths in Stereo Matching
• Iterative Geometry Encoding Volume for Stereo Matching
• CroCo v2: Improved Cross-view Completion Pre-training for Stereo Matching and Optical Flow
• Practical Stereo Matching via Cascaded Recurrent Network with Adaptive Correlation
• CMX: Cross-Modal Fusion for RGB-X Semantic Segmentation with Transformers
• DSEC: A Stereo Event Camera Dataset for Driving Scenarios
• Hierarchical Neural Architecture Search for Deep Stereo Matching
• AANet: Adaptive Aggregation Network for Efficient Stereo Matching
• Neuromorphic Visual Odometry System for Intelligent Vehicle Application with Bio-inspired Vision Sensor
• Group-wise Correlation Stereo Network
• Continuous-time Intensity Estimation Using Event Cameras
• Bounding Box Regression with Uncertainty for Accurate Object Detection
• Pyramid Stereo Matching Network
• Deformable Convolutional Networks
• End-to-End Learning of Geometry and Context for Deep Stereo Regression
• Feature Pyramid Networks for Object Detection
• Simple and Scalable Predictive Uncertainty Estimation using Deep Ensembles
• Deep Residual Learning for Image Recognition
• Dropout as a Bayesian Approximation: Representing Model Uncertainty in Deep Learning
• Fast R-CNN
• FlowNet: Learning Optical Flow with Convolutional Networks
• Stereo Anything: Unifying Zero-shot Stereo Matching with Large-Scale Mixed Data

Cited by

• Geometry-Aware Sparse Depth Sampling for High-Fidelity RGB-D Depth Completion in Robotic Systems

Related

• Deep Stereo using Adaptive Thin Volume Representation with Uncertainty Awareness
• SRFNet: Monocular Depth Estimation with Fine-grained Structure via Spatial Reliability-oriented Fusion of Frames and Events
• Event-based Stereo Depth Estimation: A Survey
• DERD-Net: Learning Depth from Event-based Ray Densities
• Event-based stereo depth estimation