跳到正文
返回论文列表
cs.CV提交于 已译

GOFish:全视场高斯溅射与网格提取

GOFish: Full Field-of-View Gaussian Splatting and Mesh Extraction

Allen Tu · Meher Gitika Karumuri · Shifan Zhu · Hailing Wang · Yen-Liang Lin · David Held · et al.

中文摘要

机器人、自动驾驶车辆和消费级360度相机使用鱼眼镜头以获得宽视场(FoV)。可仿真的数字孪生需要在整个视场内实现照片级真实感渲染,并对齐碰撞网格。然而,现有鱼眼高斯溅射方法仅能光栅化至180度,且不生成网格;而高斯表面重建方法假设训练图像为针孔模型。GOFish 是首个直接在鱼眼图像上训练、光栅化超过180度视场、并从同一场景提取对齐网格的3D高斯溅射表面重建方法。校准致密化在整个视场内恢复表面,校准Mip滤波与广角外观建模将像素模糊和镜头衰减排除在场景与网格之外。在200度FIORD和205度FullCircle基准上,GOFish 的PSNR分别比最佳先前方法提升2.28和1.97 dB;在其自身视场内也优于鱼眼渲染基线,并在FIORD和主ScanNet++集上取得最高网格F-score。高斯场景与网格共同构成无需LiDAR、可仿真的数字孪生,支持实时渲染、全视场光线追踪并提供碰撞几何。

关键要点

  1. 01现有鱼眼高斯溅射仅光栅化至180度且不生成网格,高斯表面重建假设针孔图像
  2. 02直接在鱼眼图像上训练,光栅化超过180度视场并提取对齐网格
  3. 03校准致密化、校准Mip滤波与广角外观建模处理全视场表面与镜头衰减
  4. 04FIORD和FullCircle上PSNR分别提升2.28和1.97 dB,网格F-score最高
  5. 05生成无需LiDAR、可实时渲染与光线追踪的仿真就绪数字孪生

解读

尚无解读。

原始英文摘要

arXiv:2610.37818v1 Announce Type: cross Abstract: Robots, autonomous vehicles, and consumer 360{\deg} cameras use fisheye lenses for wide fields of view (FoVs). Simulation-ready digital twins require photorealistic rendering across the full FoV and an aligned collision mesh. However, prior fisheye Gaussian Splatting methods only rasterize up to 180{\deg} and do not produce meshes, while Gaussian surface reconstruction methods assume pinhole training images. We present GOFish, the first 3D Gaussian Splatting surface reconstruction method to train directly on fisheye images, rasterize FoVs beyond 180{\deg}, and extract an aligned mesh from the same scene. Calibrated densification recovers the surface across the entire FoV, while Calibrated Mip Filtering and Wide-Angle Appearance Modeling keep pixel blur and lens falloff out of the scene and mesh. Across the 200{\deg} FIORD and 205{\deg} FullCircle benchmarks, GOFish improves PSNR over the best prior method by 2.28 and 1.97 dB, respectively. It also outperforms fisheye rendering baselines within their own FoVs and achieves the highest mesh F-score on FIORD and the main ScanNet++ set. Together, the Gaussian scene and mesh form a LiDAR-free, simulation-ready digital twin that renders in real time, ray traces the full FoV, and provides collision geometry.

同方向论文 · cs.CV

查看全部 →