EaDeblur-GS: Event assisted 3D Deblur Reconstruction with Gaussian Splatting

TL;DR

EaDeblur-GS leverages event camera data and novel neural network techniques to improve 3D deblurring reconstruction robustness and real-time performance, especially under severe motion blur conditions.

Contribution

The paper introduces EaDeblur-GS, a novel method integrating event data with Gaussian Splatting and adaptive estimation to enhance 3D deblurring capabilities.

Findings

Achieves sharp 3D reconstructions from blurry inputs.

Performs in real-time with results comparable to state-of-the-art.

Effectively handles severe motion blur and complex camera motion.

Abstract

3D deblurring reconstruction techniques have recently seen significant advancements with the development of Neural Radiance Fields (NeRF) and 3D Gaussian Splatting (3DGS). Although these techniques can recover relatively clear 3D reconstructions from blurry image inputs, they still face limitations in handling severe blurring and complex camera motion. To address these issues, we propose Event-assisted 3D Deblur Reconstruction with Gaussian Splatting (EaDeblur-GS), which integrates event camera data to enhance the robustness of 3DGS against motion blur. By employing an Adaptive Deviation Estimator (ADE) network to estimate Gaussian center deviations and using novel loss functions, EaDeblur-GS achieves sharp 3D reconstructions in real-time, demonstrating performance comparable to state-of-the-art methods.