SeaThru-NeRF: Neural Radiance Fields in Scattering Media
Deborah Levy, Amit Peleg, Naama Pearl, Dan Rosenbaum, Derya Akkaynak, Simon Korman, Tali Treibitz
Abstract
sea-thru-nerf.github.io Figure 1. NeRFs have not yet tackled scenes in which the medium strongly influences the appearances of objects, as in the case of underwater imagery. By incorporating a scattering image formation model into the NeRF rendering equations, we are able to separate the scene into 'clean' and backscatter components. Consequently, we can render photorealistic novel-views with or without the participating medium, in the latter case recovering colors as if the image was taken in clear air. Results on the Curac ¸ao scene: A RAW image (left) is brightened and white balanced (WB) for visualization, showing more detail, while areas further from the camera (top-right corner) are occluded and attenuated by severe backscatter -which is effectively removed in our restored image. Please zoom-in to observe the details.
Ask about this paper
Your agent reads all of it.
Lune indexed this paper to the last equation, along with the top-tier papers that cite it. Ask a question and the answer quotes them.
Cited by top-tier papers22
- Aleth-NeRF: Illumination Adaptive NeRF with Concealing Field AssumptionZiteng Cui, Lin Gu, Xiao Sun, Xianzheng Ma et al.AAAI 2024 · 68 citations
- AONeuS: A Neural Rendering Framework for Acoustic-Optical Sensor FusionMohamad Qadri, Kevin Zhang, Akshay Hinduja, Michael Kaess et al.SIGGRAPH 2024 · 23 citations
- Radar Fields: Frequency-Space Neural Scene Representations for FMCW RadarDavid Borts, Erich Liang, Tim Broedermann, Andrea Ramazzina et al.SIGGRAPH 2024 · 20 citations
- I2-NeRF: Learning Neural Radiance Fields Under Physically-Grounded Media InteractionsShuhong Liu, Lin Gu, Ziteng Cui, Xuangeng Chu et al.NeurIPS 2025 · 20 citations
- Neural Underwater Scene RepresentationYunkai Tang, Chengxuan Zhu, Renjie Wan, Chao Xu et al.CVPR 2024 · 19 citations
Builds on11
- Mip-NeRF 360: Unbounded Anti-Aliased Neural Radiance FieldsJonathan T. Barron, Ben Mildenhall, Dor Verbin, Pratul P. Srinivasan et al.CVPR 2022 · 1,603 citations
- Ref-NeRF: Structured View-Dependent Appearance for Neural Radiance FieldsDor Verbin, Peter Hedman, Ben Mildenhall, Todd E. Zickler et al.CVPR 2022 · 477 citations
- NeRF in the Dark: High Dynamic Range View Synthesis from Noisy Raw ImagesBen Mildenhall, Peter Hedman, Ricardo Martin-Brualla, Pratul P. Srinivasan et al.CVPR 2022 · 307 citations
- Guided Curriculum Model Adaptation and Uncertainty-Aware Evaluation for Semantic Nighttime Image SegmentationChristos Sakaridis, Dengxin Dai, Luc Van GoolICCV 2019 · 297 citations
- Deblur-NeRF: Neural Radiance Fields from Blurry ImagesLi Ma, Xiaoyu Li, Jing Liao, Qi Zhang et al.CVPR 2022 · 176 citations
Related papers
- Decoupling Scattering: Pseudo-Label Guided NeRF for Scenes with Scattering MediaMingyang Zhang, Junkang Zhang, Faming Fang, Guixu ZhangAAAI 2025 · 4 citations
- AtlantisGS: Underwater Sparse-View Scene Reconstruction via Gaussian SplattingJingjun Yi, Qi Bi, Hao Zheng, Huimin Huang et al.ACM MM 2025 · 1 citation
- NeuroPump: Simultaneous Geometric and Color Rectification for Underwater ImagesYue Guo, Haoxiang Liao, Haibin Ling, Bingyao HuangACM MM 2025 · 1 citation
- NeRFrac: Neural Radiance Fields through Refractive SurfaceYifan Zhan, Shohei Nobuhara, Ko Nishino, Yinqiang ZhengICCV 2023 · 18 citations
- Relative Illumination Fields: Learning Medium and Light Independent Underwater ScenesMengkun She, Felix Seegräber, David Nakath, Patricia Schöntag et al.ICCV 2025
