Geometric-Photometric Event-based 3D Gaussian Ray Tracing

arXiv cs.RO / 4/2/2026

💬 OpinionSignals & Early TrendsIdeas & Deep AnalysisModels & Research

共有:

Key Points

The paper introduces GPERT, a framework for event-based 3D Gaussian Splatting that targets the accuracy vs. temporal-resolution trade-off inherent to sparse event streams.
It decouples rendering into two components—event-by-event geometry (depth) via ray tracing and snapshot-based radiance (intensity) using warped-event images—so fine-grained timing can be exploited without conflating geometry and appearance.
Experiments report state-of-the-art results on real-world datasets and competitive performance on a synthetic dataset.
The method claims to work without prior information such as pretrained image reconstruction models or COLMAP-based initialization, and it is flexible in the number of events used.
The approach is reported to produce sharp reconstructions at scene edges while enabling fast training, with the authors positioning the work as a step toward better understanding how event sparsity affects 3D reconstruction.

Abstract

Event cameras offer a high temporal resolution over traditional frame-based cameras, which makes them suitable for motion and structure estimation. However, it has been unclear how event-based 3D Gaussian Splatting (3DGS) approaches could leverage fine-grained temporal information of sparse events. This work proposes GPERT, a framework to address the trade-off between accuracy and temporal resolution in event-based 3DGS. Our key idea is to decouple the rendering into two branches: event-by-event geometry (depth) rendering and snapshot-based radiance (intensity) rendering, by using ray-tracing and the image of warped events. The extensive evaluation shows that our method achieves state-of-the-art performance on the real-world datasets and competitive performance on the synthetic dataset. Also, the proposed method works without prior information (e.g., pretrained image reconstruction models) or COLMAP-based initialization, is more flexible in the event selection number, and achieves sharp reconstruction on scene edges with fast training time. We hope that this work deepens our understanding of the sparse nature of events for 3D reconstruction. https://github.com/e3ai/gpert