# BenthicFlow: Generating Extensible Underwater Environments via Flow Matching

arXiv: https://arxiv.org/abs/2608.23173 · HTML: https://mlsanity.com/paper/2608.23173/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Inverse problems and image restoration, one of 14 directions in Diffusion models.
- Builds on 26 of its 57 references that are in the map.
- First posted on arXiv 24 Aug 2026.

## Placed in

- Diffusion models / [Inverse problems and image restoration](https://mlsanity.com/diffusion/inverse-problems-and-image-restoration/)

## Builds on (most cited first, 26 of 26)

- [U-Net: Convolutional Networks for Biomedical Image Segmentation](https://mlsanity.com/paper/1505.04597/) (2015)
- [High-Resolution Image Synthesis with Latent Diffusion Models](https://mlsanity.com/paper/2112.10752/) (2021)
- [The Unreasonable Effectiveness of Deep Features as a Perceptual Metric](https://mlsanity.com/paper/1801.03924/) (2018)
- [Scalable Diffusion Models with Transformers](https://mlsanity.com/paper/2212.09748/) (2022)
- [Classifier-Free Diffusion Guidance](https://mlsanity.com/paper/2207.12598/) (2022)
- [Flow Matching for Generative Modeling](https://mlsanity.com/paper/2210.02747/) (2022)
- [Rethinking the Inception Architecture for Computer Vision](https://mlsanity.com/paper/1512.00567/) (2015)
- [Scaling Rectified Flow Transformers for High-Resolution Image Synthesis](https://mlsanity.com/paper/2403.03206/) (2024)
- [Flow Straight and Fast: Learning to Generate and Transfer Data with Rectified Flow](https://mlsanity.com/paper/2209.03003/) (2022)
- [Taming Transformers for High-Resolution Image Synthesis](https://mlsanity.com/paper/2012.09841/) (2020)
- [Depth Anything V2](https://mlsanity.com/paper/2406.09414/) (2024)
- [2D Gaussian Splatting for Geometrically Accurate Radiance Fields](https://mlsanity.com/paper/2403.17888/) (2024)
- [Building Normalizing Flows with Stochastic Interpolants](https://mlsanity.com/paper/2209.15571/) (2022)
- [MultiDiffusion: Fusing Diffusion Paths for Controlled Image Generation](https://mlsanity.com/paper/2302.08113/) (2023)
- [Diffusion Transformers with Representation Autoencoders](https://mlsanity.com/paper/2510.11690/) (2025)
- [Fixup Initialization: Residual Learning Without Normalization](https://mlsanity.com/paper/1901.09321/) (2019)
- [SeaThru-NeRF: Neural Radiance Fields in Scattering Media](https://mlsanity.com/paper/2304.07743/) (2023)
- [Contrastive Semi-supervised Learning for Underwater Image Restoration via Reliable Bank](https://mlsanity.com/paper/2303.09101/) (2023)
- [Mixture of Diffusers for scene composition and high resolution image generation](https://mlsanity.com/paper/2302.02412/) (2023)
- [WaterSplatting: Fast Underwater 3D Scene Reconstruction Using Gaussian Splatting](https://mlsanity.com/paper/2408.08206/) (2024)
- [Underwater Light Field Retention : Neural Rendering for Underwater Imaging](https://mlsanity.com/paper/2203.11006/) (2022)
- [Real-Time Dense 3D Mapping of Underwater Environments](https://mlsanity.com/paper/2304.02704/) (2023)
- [OceanSim: A GPU-Accelerated Underwater Robot Perception Simulation Framework](https://mlsanity.com/paper/2503.01074/) (2025)
- [Combining Photogrammetric Computer Vision and Semantic Segmentation for Fine-grained Understanding of Coral Reef Growth under Climate Change](https://mlsanity.com/paper/2212.04132/) (2022)
- [DPF-Net: Physical Imaging Model Embedded Data-Driven Underwater Image Enhancement](https://mlsanity.com/paper/2503.12470/) (2025)
- [Infinite Leagues Under the Sea: Photorealistic 3D Underwater Terrain Generation by Latent Fractal Diffusion Models](https://mlsanity.com/paper/2503.06784/) (2025)

## Abstract (arXiv)

Computer vision applications for 3D scene understanding in underwater environments remain challenging due to the lack of high-quality 3D data and the inability of surface-trained models to generalize to underwater scenes. To address this challenge, an emerging trend is to employ generative models to close the data domain gap. However, existing methods assemble large scenes by stitching independently generated tiles post hoc with separately trained models, while demonstrating heterogeneous landscapes only within individual survey sites. We introduce BenthicFlow, a unified framework based on a single conditional flow-matching model that jointly generates aligned textures and depth maps. A MultiDiffusion-inspired sampling procedure reconciles overlapping windows throughout the generative trajectory, enabling spatially extensible RGBD mosaics without a separate stitching model. The generated mosaics are subsequently lifted into explicit 3D benthic environments using surface-aligned Gaussian surfels. Experiments across geographically distinct survey sites demonstrate that BenthicFlow preserves site-specific appearance while generating coherent, large-scale 3D scenes that closely match the target distributions. Code and trained models are available at https://github.com/jacomof/BenthicFlow.