SaENeRF: Suppressing Artifacts in Event-based Neural Radiance Fields

2025-04-23

AI Summary

Addressing the concurrent presence of geometric distortion and photometric artifacts in event-camera data reconstruction, this paper proposes a robust Event-NeRF framework. Methodologically, it introduces (1) a novel radiance-change normalization mechanism based on polarity-aware event accumulation to mitigate inherent event-stream noise, and (2) a dual-objective regularization loss that jointly suppresses sub-threshold artifacts and enhances brightness contrast for non-zero events. The method enables end-to-end, self-supervised optimization from raw event streams to neural radiance fields, incorporating polarity-guided radiance modeling. Experiments demonstrate substantial improvements in reconstruction density, geometric consistency, and photometric fidelity for static scenes. Quantitatively, our approach achieves superior PSNR and SSIM scores; qualitatively, it outperforms existing state-of-the-art methods in visual quality and structural accuracy.

Technology Category

• Computer Vision: Adversarial Attacks & Robustness
• Intelligent Robots: Multimodal Perception & Sensor Fusion
• Search and Optimization: Learning to Search

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

Event cameras are neuromorphic vision sensors that asynchronously capture changes in logarithmic brightness changes, offering significant advantages such as low latency, low power consumption, low bandwidth, and high dynamic range. While these characteristics make them ideal for high-speed scenarios, reconstructing geometrically consistent and photometrically accurate 3D representations from event data remains fundamentally challenging. Current event-based Neural Radiance Fields (NeRF) methods partially address these challenges but suffer from persistent artifacts caused by aggressive network learning in early stages and the inherent noise of event cameras. To overcome these limitations, we present SaENeRF, a novel self-supervised framework that effectively suppresses artifacts and enables 3D-consistent, dense, and photorealistic NeRF reconstruction of static scenes solely from event streams. Our approach normalizes predicted radiance variations based on accumulated event polarities, facilitating progressive and rapid learning for scene representation construction. Additionally, we introduce regularization losses specifically designed to suppress artifacts in regions where photometric changes fall below the event threshold and simultaneously enhance the light intensity difference of non-zero events, thereby improving the visual fidelity of the reconstructed scene. Extensive qualitative and quantitative experiments demonstrate that our method significantly reduces artifacts and achieves superior reconstruction quality compared to existing methods. The code is available at https://github.com/Mr-firework/SaENeRF.

Problem

• Suppressing artifacts in event-based NeRF reconstructions
• Improving 3D consistency and photorealism from event data
• Addressing noise and early-stage overfitting in event NeRF training

Innovation

• Self-supervised framework suppressing NeRF artifacts
• Normalizes radiance via accumulated event polarities
• Regularization losses enhance photometric fidelity