AsyncEvGS: Asynchronous Event-Assisted Gaussian Splatting for Handheld Motion-Blurred Scenes

2026-05-07

AI Summary

This work addresses the significant performance degradation of existing 3D reconstruction methods when applied to severely motion-blurred images captured by handheld devices, as well as the limitations of event camera–assisted approaches due to low spatial resolution and stringent synchronization requirements. We propose the first high-resolution, asynchronous RGB-event fusion framework tailored for standard handheld devices, which leverages event data to recover sharp images and introduces a cross-domain pose estimation module based on the Visual Geometry Transformer to provide robust initialization for 3D Gaussian Splatting. To further mitigate the ill-posedness of conventional losses under motion blur, we incorporate a structure-driven event loss and a view-consistency regularizer. Evaluated on our newly curated high-resolution asynchronous dataset, AsyncEv-Deblur, and established benchmarks, our method substantially improves reconstruction robustness and fidelity.

Abstract

3D reconstruction methods such as 3D Gaussian Splatting (3DGS) and Neural Radiance Fields (NeRF) achieve impressive photorealism but fail when input images suffer from severe motion blur. While event cameras provide high-temporal-resolution motion cues, existing event-assisted approaches rely on low-resolution sensors and strict synchronization, limiting their practicality for handheld 3D capture on common devices, such as smartphones. We introduce a flexible, high-resolution asynchronous RGB-Event dual-camera system and a corresponding reconstruction framework. Our approach first reconstructs sharp images from the event data and then employs a cross-domain pose estimation module based on the Visual Geometry Transformer (VGGT) to obtain robust initialization for 3DGS. During optimization, we employ a structure-driven event loss and view-specific consistency regularizers to mitigate the ill-posed behavior of traditional event losses and deblurring losses, ensuring both stable and high-fidelity reconstruction. We further contribute AsyncEv-Deblur, a new high-resolution RGB-Event dataset captured with our asynchronous system. Experiments demonstrate that our method achieves state-of-the-art performance on both our challenging dataset and existing benchmarks, substantially improving reconstruction robustness under severe motion blur. Project page: https://openimaginglab.github.io/AsyncEvGS/

Problem

• motion blur
• 3D reconstruction
• event camera
• asynchronous sensing
• handheld capture

Innovation

• asynchronous event camera
• Gaussian Splatting
• motion deblurring
• cross-domain pose estimation
• structure-driven event loss