EventNeuS: 3D Mesh Reconstruction from a Single Event Camera

2026-02-03

AI Summary

Existing event camera–based 3D reconstruction methods struggle to achieve high-fidelity dense mesh reconstruction. This work proposes a self-supervised neural implicit representation that, for the first time, jointly models a signed distance function and a density field, while incorporating spherical harmonics encoding to capture view-dependent effects. The method enables high-quality 3D reconstruction from monocular color event streams alone. Experimental results demonstrate significant improvements in reconstruction accuracy, with reductions of 34% in Chamfer distance and 31% in mean absolute error compared to the current state-of-the-art approach.

Technology Category

• Computer Vision: 3D Computer Vision
• Machine Learning: Representation Learning
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Search and Retrieval-Augmented AI: Web query analysis, representation and understanding
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation

Abstract

Event cameras offer a considerable alternative to RGB cameras in many scenarios. While there are recent works on event-based novel-view synthesis, dense 3D mesh reconstruction remains scarcely explored and existing event-based techniques are severely limited in their 3D reconstruction accuracy. To address this limitation, we present EventNeuS, a self-supervised neural model for learning 3D representations from monocular colour event streams. Our approach, for the first time, combines 3D signed distance function and density field learning with event-based supervision. Furthermore, we introduce spherical harmonics encodings into our model for enhanced handling of view-dependent effects. EventNeuS outperforms existing approaches by a significant margin, achieving 34% lower Chamfer distance and 31% lower mean absolute error on average compared to the best previous method.

Problem

• event camera
• 3D mesh reconstruction
• dense reconstruction
• 3D reconstruction accuracy

Innovation

• event camera
• 3D mesh reconstruction
• signed distance function
• neural radiance fields
• spherical harmonics