EventNeuS: 3D Mesh Reconstruction from a Single Event Camera

TL;DR

EventNeuS is a novel self-supervised neural model that reconstructs accurate 3D meshes from single event camera streams, combining signed distance functions and spherical harmonics for improved performance.

Contribution

It introduces a new approach that integrates 3D signed distance functions with event-based supervision and view-dependent encoding, advancing 3D reconstruction from event data.

Findings

Achieves 34% lower Chamfer distance than previous methods.

Attains 31% lower mean absolute error in 3D reconstruction.

Outperforms existing event-based 3D reconstruction techniques significantly.

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.