# Lumos3D: A Single-Forward Framework for Low-Light 3D Scene Restoration

arXiv: https://arxiv.org/abs/2511.09818 · HTML: https://mlsanity.com/paper/2511.09818/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 23 of its 24 references that are in the map.
- First posted on arXiv 12 Nov 2025.

## Builds on (most cited first, 23 of 23)

- [Very Deep Convolutional Networks for Large-Scale Image Recognition](https://mlsanity.com/paper/1409.1556/) (2014)
- [The Unreasonable Effectiveness of Deep Features as a Perceptual Metric](https://mlsanity.com/paper/1801.03924/) (2018)
- [3D Gaussian Splatting for Real-Time Radiance Field Rendering](https://mlsanity.com/paper/2308.04079/) (2023)
- [DUSt3R: Geometric 3D Vision Made Easy](https://mlsanity.com/paper/2312.14132/) (2023)
- [VGGT: Visual Geometry Grounded Transformer](https://mlsanity.com/paper/2503.11651/) (2025)
- [Depth Anything V2](https://mlsanity.com/paper/2406.09414/) (2024)
- [Grounding Image Matching in 3D with MASt3R](https://mlsanity.com/paper/2406.09756/) (2024)
- [Continuous 3D Perception Model with Persistent State](https://mlsanity.com/paper/2501.12387/) (2025)
- [DL3DV-10K: A Large-Scale Scene Dataset for Deep Learning-based 3D Vision](https://mlsanity.com/paper/2312.16256/) (2023)
- [3D Reconstruction with Spatial Memory](https://mlsanity.com/paper/2408.16061/) (2024)
- [AnySplat: Feed-forward 3D Gaussian Splatting from Unconstrained Views](https://mlsanity.com/paper/2505.23716/) (2025)
- [DeMoN: Depth and Motion Network for Learning Monocular Stereo](https://mlsanity.com/paper/1612.02401/) (2016)
- [DeepV2D: Video to Depth with Differentiable Structure from Motion](https://mlsanity.com/paper/1812.04605/) (2018)
- [MUSt3R: Multi-view Network for Stereo 3D Reconstruction](https://mlsanity.com/paper/2503.01661/) (2025)
- [DeepTAM: Deep Tracking and Mapping](https://mlsanity.com/paper/1808.01900/) (2018)
- [Multi-view 3D Reconstruction with Transformer](https://mlsanity.com/paper/2103.12957/) (2021)
- [Lighting up NeRF via Unsupervised Decomposition and Enhancement](https://mlsanity.com/paper/2307.10664/) (2023)
- [Aleth-NeRF: Illumination Adaptive NeRF with Concealing Field Assumption](https://mlsanity.com/paper/2312.09093/) (2023)
- [Bilateral Guided Radiance Field Processing](https://mlsanity.com/paper/2406.00448/) (2024)
- [Luminance-GS: Adapting 3D Gaussian Splatting to Challenging Lighting Conditions with View-Adaptive Curve Adjustment](https://mlsanity.com/paper/2504.01503/) (2025)
- [LuSh-NeRF: Lighting up and Sharpening NeRFs for Low-light Scenes](https://mlsanity.com/paper/2411.06757/) (2024)
- [R3eVision: A Survey on Robust Rendering, Restoration, and Enhancement for 3D Low-Level Vision](https://mlsanity.com/paper/2506.16262/) (2025)
- [Stylos: Multi-View 3D Stylization with Single-Forward Gaussian Splatting](https://mlsanity.com/paper/2509.26455/) (2025)

## Built on by (most cited first, 1 of 1)

- [RealX3D: A Physically-Degraded 3D Benchmark for Multi-view Visual Restoration and Reconstruction](https://mlsanity.com/paper/2512.23437/) (2025)

## Abstract (arXiv)

Restoring 3D scenes with low-light conditions is challenging, and most existing methods depend on precomputed camera poses and scene-specific optimization, which greatly restricts their application to real-world scenarios. To overcome these limitations, we propose Lumos3D, a pose-free single-forward framework for 3D low-light scene restoration. First, we develop a cross-illumination distillation scheme, where a frozen teacher network takes normal-light ground truth images as input to distill accurate geometric information to the student model. Second, we define a Lumos loss to improve the restoration quality of the reconstructed 3D Gaussian space. Trained on a single dataset, Lumos3D performs inference in a purely feed-forward manner, directly restoring illumination and structure from unposed, low-light multi-view images without any per-scene training or optimization. Experiments on real-world datasets demonstrate that Lumos3D achieves competitive restoration results compared to scene-specific methods. Our codes are available at https://github.com/HanzhouLiu/Lumos3D.