# From Chaos to Clarity: 3DGS in the Dark

arXiv: https://arxiv.org/abs/2406.08300 · HTML: https://mlsanity.com/paper/2406.08300/

## Key facts

- Cited by 15 papers in the map (data through 14 Oct 2026).
- Builds on 29 of its 40 references that are in the map.
- First posted on arXiv 12 Jun 2024.

## Builds on (most cited first, 29 of 29)

- [3D Gaussian Splatting for Real-Time Radiance Field Rendering](https://mlsanity.com/paper/2308.04079/) (2023)
- [Instant Neural Graphics Primitives with a Multiresolution Hash Encoding](https://mlsanity.com/paper/2201.05989/) (2022)
- [Plenoxels: Radiance Fields without Neural Networks](https://mlsanity.com/paper/2112.05131/) (2021)
- [Magic3D: High-Resolution Text-to-3D Content Creation](https://mlsanity.com/paper/2211.10440/) (2022)
- [Deep Image Prior](https://mlsanity.com/paper/1711.10925/) (2017)
- [PlenOctrees for Real-time Rendering of Neural Radiance Fields](https://mlsanity.com/paper/2103.14024/) (2021)
- [Scaffold-GS: Structured 3D Gaussians for View-Adaptive Rendering](https://mlsanity.com/paper/2312.00109/) (2023)
- [KiloNeRF: Speeding up Neural Radiance Fields with Thousands of Tiny MLPs](https://mlsanity.com/paper/2103.13744/) (2021)
- [FastNeRF: High-Fidelity Neural Rendering at 200FPS](https://mlsanity.com/paper/2103.10380/) (2021)
- [Latent-NeRF for Shape-Guided Generation of 3D Shapes and Textures](https://mlsanity.com/paper/2211.07600/) (2022)
- [Baking Neural Radiance Fields for Real-Time View Synthesis](https://mlsanity.com/paper/2103.14645/) (2021)
- [Learning to See in the Dark](https://mlsanity.com/paper/1805.01934/) (2018)
- [Neural Geometric Level of Detail: Real-time Rendering with Implicit 3D Shapes](https://mlsanity.com/paper/2101.10994/) (2021)
- [NeRF in the Dark: High Dynamic Range View Synthesis from Noisy Raw Images](https://mlsanity.com/paper/2111.13679/) (2021)
- [FSGS: Real-Time Few-shot View Synthesis using Gaussian Splatting](https://mlsanity.com/paper/2312.00451/) (2023)
- [GRAM: Generative Radiance Manifolds for 3D-Aware Image Generation](https://mlsanity.com/paper/2112.08867/) (2021)
- [Neighbor2Neighbor: Self-Supervised Denoising from Single Noisy Images](https://mlsanity.com/paper/2101.02824/) (2021)
- [A Physics-based Noise Formation Model for Extreme Low-light Raw Denoising](https://mlsanity.com/paper/2003.12751/) (2020)
- [Variable Bitrate Neural Fields](https://mlsanity.com/paper/2206.07707/) (2022)
- [Supervised Raw Video Denoising with a Benchmark Dataset on Dynamic Scenes](https://mlsanity.com/paper/2003.14013/) (2020)
- [READ: Large-Scale Neural Scene Rendering for Autonomous Driving](https://mlsanity.com/paper/2205.05509/) (2022)
- [Instant-3D: Instant Neural Radiance Field Training Towards On-Device AR/VR 3D Reconstruction](https://mlsanity.com/paper/2304.12467/) (2023)
- [LG-BPN: Local and Global Blind-Patch Network for Self-Supervised Real-World Denoising](https://mlsanity.com/paper/2304.00534/) (2023)
- [RT-NeRF: Real-Time On-Device Neural Radiance Fields Towards Immersive AR/VR Rendering](https://mlsanity.com/paper/2212.01120/) (2022)
- [Synthetic Defocus and Look-Ahead Autofocus for Casual Videography](https://mlsanity.com/paper/1905.06326/) (2019)
- [Disorder-invariant Implicit Neural Representation](https://mlsanity.com/paper/2304.00837/) (2023)
- [Learnability Enhancement for Low-light Raw Denoising: Where Paired Real Data Meets Noise Modeling](https://mlsanity.com/paper/2207.06103/) (2022)
- [Efficient Visual Computing with Camera RAW Snapshots](https://mlsanity.com/paper/2212.07778/) (2022)
- [Regulating Intermediate 3D Features for Vision-Centric Autonomous Driving](https://mlsanity.com/paper/2312.11837/) (2023)

## Built on by (most cited first, 15 of 15)

- [RealX3D: A Physically-Degraded 3D Benchmark for Multi-view Visual Restoration and Reconstruction](https://mlsanity.com/paper/2512.23437/) (2025)
- [LL-Gaussian: Low-Light Scene Reconstruction and Enhancement via Gaussian Splatting for Novel View Synthesis](https://mlsanity.com/paper/2504.10331/) (2025)
- [RobustGS: Unified Boosting of Feedforward 3D Gaussian Splatting under Low-Quality Conditions](https://mlsanity.com/paper/2508.03077/) (2025)
- [GaussHDR: High Dynamic Range Gaussian Splatting via Learning Unified 3D and 2D Local Tone Mapping](https://mlsanity.com/paper/2503.10143/) (2025)
- [Unifying Color and Lightness Correction with View-Adaptive Curve Adjustment for Robust 3D Novel View Synthesis](https://mlsanity.com/paper/2602.18322/) (2026)
- [RF4D:Neural Radar Fields for Novel View Synthesis in Outdoor Dynamic Scenes](https://mlsanity.com/paper/2505.20967/) (2025)
- [RAWild: Sensor-Agnostic RAW Object Detection via Physics-Guided Curve and Grid Modeling](https://mlsanity.com/paper/2605.05941/) (2026)
- [Time of the Flight of the Gaussians: Optimizing Depth Indirectly in Dynamic Radiance Fields](https://mlsanity.com/paper/2505.05356/) (2025)
- [Mono4DGS-HDR: High Dynamic Range 4D Gaussian Splatting from Alternating-exposure Monocular Videos](https://mlsanity.com/paper/2510.18489/) (2025)
- [Naka-GS: Naka-Guided Chroma Correction and Point-Prior Compaction for Low-Light 3D Gaussian Splatting](https://mlsanity.com/paper/2604.11142/) (2026)
- [Physically Inspired Gaussian Splatting for HDR Novel View Synthesis](https://mlsanity.com/paper/2603.28020/) (2026)
- [PanDORA: Casual HDR Radiance Acquisition of Indoor Scenes for Image-based Lighting](https://mlsanity.com/paper/2407.06150/) (2024)
- [InstantHDR: Single-forward Gaussian Splatting Initialization for HDR 3D Reconstruction](https://mlsanity.com/paper/2603.11298/) (2026)
- [JSGS: JPEG State-Guided Supervision for 3D Gaussian Splatting from Mixed-Quality Views](https://mlsanity.com/paper/2608.08659/) (2026)
- [FujinSplat: Seeing Through Smoke with RAW-Domain Gaussian Splatting](https://mlsanity.com/paper/2609.06017/) (2026)

## Abstract (arXiv)

Novel view synthesis from raw images provides superior high dynamic range (HDR) information compared to reconstructions from low dynamic range RGB images. However, the inherent noise in unprocessed raw images compromises the accuracy of 3D scene representation. Our study reveals that 3D Gaussian Splatting (3DGS) is particularly susceptible to this noise, leading to numerous elongated Gaussian shapes that overfit the noise, thereby significantly degrading reconstruction quality and reducing inference speed, especially in scenarios with limited views. To address these issues, we introduce a novel self-supervised learning framework designed to reconstruct HDR 3DGS from a limited number of noisy raw images. This framework enhances 3DGS by integrating a noise extractor and employing a noise-robust reconstruction loss that leverages a noise distribution prior. Experimental results show that our method outperforms LDR/HDR 3DGS and previous state-of-the-art (SOTA) self-supervised and supervised pre-trained models in both reconstruction quality and inference speed on the RawNeRF dataset across a broad range of training views. Code can be found in urlhttps://lizhihao6.github.io/Raw3DGS.