GlowGS: Generative Semantic Feature Learning for 3D Gaussian Splatting in Nighttime Glow Scenes

2026-05-22

AI Summary

Existing 3D Gaussian splatting methods struggle to achieve high-quality novel view synthesis in nighttime luminous scenes due to the absence of structural priors such as texture and edges. This work proposes, for the first time, integrating generative semantic features into the 3D Gaussian splatting framework by leveraging diffusion models and vision foundation models. It introduces a label-free semantic feature generation mechanism and a novel-view semantic learning strategy that implicitly impose structural constraints during reconstruction. The proposed approach significantly enhances semantic consistency and visual fidelity in nighttime scene rendering while effectively suppressing artifacts, outperforming current state-of-the-art methods under fully unsupervised conditions.

Abstract

Existing 3DGS methods effectively render high-quality novel views in clear-day scenes. However, they struggle with night scenes, particularly in glow regions, due to the lack of structural features such as textures and edges, which are key cues for splatting-based reconstruction. To address this problem, we leverage a diffusion model and a Vision Foundation Model (VFM) to compensate for missing structural cues. Our method consists of two key novel ideas: semantic feature generation and novel-view semantic learning. First, semantic feature generation produces high-quality semantic features as implicit structural cues for novel views. Specifically, a diffusion model synthesizes novel views with unknown camera poses from training views, while a VFM evaluates their quality. Once high-quality novel views are identified, the VFM extracts robust features to construct the semantic feature bank. Second, novel-view semantic learning enables 3DGS to optimize rendered novel views without requiring ground truth. It achieves this by extracting semantic features from a rendered novel view, searching the feature bank for the most similar features, and minimizing their distance. This process enforces implicit structural constraints, ensuring semantically coherent, artifact-free rendered views. Extensive experiments demonstrate the effectiveness of our GlowGS in generating semantically accurate 3D views, showing significant improvements over existing methods.

Problem

• 3D Gaussian Splatting
• nighttime glow scenes
• structural features
• novel view synthesis
• semantic coherence

Innovation

• 3D Gaussian Splatting
• semantic feature generation
• diffusion model
• Vision Foundation Model
• nighttime glow scenes