Low-light 3D scene capture improved with noise-aware gaussian splatting
NOVA-GS: Noise-Aware View-Consistent Gaussian Splatting for Low-Light Novel View Synthesis
Computer Vision and Pattern Recognition
Summary
Capturing 3D scenes in very dark places is hard because camera images have a lot of noise and inconsistent lighting, which makes 3D reconstruction unstable. The authors created NOVA-GS, a method that fixes lighting and removes noise all at once while building a 3D scene from the dark images without needing a clear reference. Their approach uses smart steps to estimate camera positions, improve image quality, and ensure consistency between different views, resulting in better and more reliable 3D models in low light conditions.
What this means in practice
- •For augmented reality developers: Create 3D scenes from low-light camera footage that maintain geometric accuracy and color consistency without needing well-lit reference images.
- •For robotics engineers: Improve robot perception in dark environments by generating reliable 3D scene reconstructions directly from noisy low-light sensor data.
Authors
Shaurya Pavan A, Vemunuri Divya Madhuri, Yash Pradeep Gawande, Kaushik Mitra
Abstract
Reconstructing 3D scenes under real-world low-light conditions remains challenging due to severe sensor noise, low signal-to-noise ratios, and degraded photometric consistency, which destabilize geometry estimation and novel view synthesis. Existing approaches often rely on well-lit reference data for reliable Structure-from-Motion (SfM) initialization under degraded inputs or apply per-view enhancement methods that introduce cross-view inconsistencies. To address these limitations, we propose \textbf{NOVA-GS}, a unified noise-aware framework for low-light 3D Gaussian Splatting that subsumes enhancement, denoising, and geometry optimization within a single process. Our method leverages VGGT-based feed-forward estimation to obtain robust camera poses and geometry directly from degraded inputs, eliminating the need for SfM. Building on this initialization, NOVA-GS integrates three coupled components: a structure-aware enhancement module for exposure correction, a self-supervised denoising module with blind-spot masking for pseudo-supervision, and a consistency-driven Gaussian Splatting optimization enforcing cross-view geometric coherence. We further introduce a noise-guided spherical harmonic regularization to suppress view-dependent artifacts in noisy regions. Extensive experiments on diverse real-world low-light datasets demonstrate improved geometric fidelity, color consistency, and robustness without requiring paired supervision or well-lit references. https://shaurya2524.github.io/nova-gs/