Antoine Guédon

arXiv 2026

From Blobs to Spokes: High-Fidelity Surface Reconstruction via Oriented Gaussians

Diego Gomez*, Antoine Guédon*, Nissim Maruani, Bingchen Gong, Maks Ovsjanikov

We interpret 3D Gaussians as stochastic oriented surface elements and derive closed-form vacancy and normal fields, enabling fast, watertight, and compact mesh extraction of full 3D scenes. Our method recovers extremely thin structures such as bicycle spokes while producing meshes with significantly fewer vertices than concurrent works.

*Both authors contributed equally to the paper.

Paper Code

@misc{gomez2026blobsspokeshighfidelitysurface,
    title={From Blobs to Spokes: High-Fidelity Surface Reconstruction via Oriented Gaussians},
    author={Gomez, Diego and Gu{\'e}don, Antoine and Maruani, Nissim and Gong, Bingchen and Ovsjanikov, Maks},
    year={2026},
    eprint={2604.07337},
    archivePrefix={arXiv},
    primaryClass={cs.CV},
    url={https://arxiv.org/abs/2604.07337},
}

CVPR 2026 · Oral

MAGICIAN: Efficient Long-Term Planning with Imagined Gaussians for Active Mapping

Shiyao Li, Antoine Guédon, Shizhe Chen, Vincent Lepetit

We introduce MAGICIAN, a novel framework capable of generating long-horizon trajectories for active 3D mapping. We propose Imagined Gaussians, derived from a neural occupancy field, to enable efficient and reliable coverage gain prediction from new viewpoints in unknown scenes, supporting feasible long-term planning with tree search.

Paper Code

@inproceedings{li2026magician,
    title={MAGICIAN: Efficient Long-Term Planning with Imagined Gaussians for Active Mapping},
    author={Li, Shiyao and Gu{\'e}don, Antoine and Chen, Shizhe and Lepetit, Vincent},
    booktitle={CVPR},
    year={2026},
}

SIGGRAPH Asia 2025 · TOG

MILo: Mesh-In-the-Loop Gaussian Splatting for Detailed and Efficient Surface Reconstruction

Antoine Guédon*, Diego Gomez*, Nissim Maruani, Bingchen Gong, George Drettakis, Maks Ovsjanikov

We propose a novel differentiable mesh extraction framework that operates during the optimization of 3DGS representations. At every training iteration, we differentiably extract a mesh only from Gaussian parameters. This enables gradient flow from the mesh to Gaussians, and guides Gaussians toward configurations better suited for surface reconstruction, resulting in better meshes with significantly fewer vertices.

*Both authors contributed equally to the paper.

Paper Code

@article{guedon2025milo,
    author       = {Gu{\'e}don, Antoine and Gomez, Diego and Maruani, Nissim and Gong, Bingchen and Drettakis, George and Ovsjanikov, Maks},
    title        = {MILo: Mesh-In-the-Loop Gaussian Splatting for Detailed and Efficient Surface Reconstruction},
    journal      = {ACM Transactions on Graphics},
    year         = {2025},
    url          = {https://anttwo.github.io/milo/}
}

CVPR 2025 · Highlight

MAtCha Gaussians: Atlas of Charts for High-Quality Geometry and Photorealism From Sparse Views

Antoine Guédon, Tomoki Ichikawa, Kohei Yamashita, Ko Nishino

We propose a novel surface representation for reconstructing high-quality 3D meshes with photorealistic rendering from sparse-view images. Our key idea is to model the underlying scene geometry as an Atlas of Charts which we render with 2D Gaussian surfels. Combined with a sparse-view SfM model like MASt3R-SfM, MAtCha can recover sharp meshes of both foreground and background objects within minutes from a few unposed RGB images.

Paper Code

@article{guedon2025matcha,
    title={MAtCha Gaussians: Atlas of Charts for High-Quality Geometry and Photorealism From Sparse Views},
    author={Gu{\'e}don, Antoine and Ichikawa, Tomoki and Yamashita, Kohei and Nishino, Ko},
    journal={CVPR},
    year={2025},
}

ECCV 2024 · Oral

Gaussian Frosting: Editable Complex Radiance Fields with Real-Time Rendering

Antoine Guédon, Vincent Lepetit

We propose to represent surfaces by a mesh covered with a "Frosting" layer of varying thickness and made of 3D Gaussians. This representation captures both complex volumetric effects created by fuzzy materials such as hair or grass as well as flat surfaces. Built from RGB images only, it can be rendered in real-time and animated using traditional animation tools.

Paper Code Blender Add-On

@article{guedon2024frosting,
    title={Gaussian Frosting: Editable Complex Radiance Fields with Real-Time Rendering},
    author={Gu{\'e}don, Antoine and Lepetit, Vincent},
    journal={ECCV},
    year={2024}
}

CVPR 2024

SuGaR: Surface-Aligned Gaussian Splatting for Efficient 3D Mesh Reconstruction and High-Quality Mesh Rendering

Antoine Guédon, Vincent Lepetit

We introduce a method that extracts accurate and editable meshes from 3D Gaussian Splatting representations within minutes on a single GPU. This enables easy editing, sculpting, rigging, animating, or relighting of the Gaussians using traditional softwares (Blender, Unity, Unreal Engine, etc.) by manipulating the mesh instead of the Gaussians themselves.

Paper Code Blender Add-On

@article{guedon2023sugar,
    title={SuGaR: Surface-Aligned Gaussian Splatting for Efficient 3D Mesh Reconstruction and High-Quality Mesh Rendering},
    author={Gu{\'e}don, Antoine and Lepetit, Vincent},
    journal={CVPR},
    year={2024}
}

CVPR 2023

MACARONS: Mapping And Coverage Anticipation with RGB ONline Self-supervision

Antoine Guédon, Tom Monnier, Pascal Monasse, Vincent Lepetit

We introduce a novel method that simultaneously learns to explore new large environments and to reconstruct them in 3D from color images only, in a self-supervised fashion.

Paper Code

@inproceedings{guedon2023macarons,
    title={{MACARONS: Mapping And Coverage Anticipation with RGB ONline Self-supervision}},
    author={Gu{\'e}don, Antoine and Monnier, Tom and Monasse, Pascal and Lepetit, Vincent},
    booktitle={{CVPR}},
    year={2023},
}

NeurIPS 2022 · Spotlight

SCONE: Surface Coverage Optimization in uNknown Environments by Volumetric Integration

Antoine Guédon, Pascal Monasse, Vincent Lepetit

We introduce a novel approach to solve the Next Best View problem for dense 3D reconstruction in unknown environments. Contrary to other learning-based methods, our approach scales to large 3D scenes and handles completely free camera motion at inference.

Paper Code

@inproceedings{guedon2022scone,
    title={{SCONE: Surface Coverage Optimization in Unknown Environments by Volumetric Integration}},
    author={Gu{\'e}don, Antoine and Monasse, Pascal and Lepetit, Vincent},
    booktitle={{Advances in Neural Information Processing Systems}},
    year={2022},
}

Antoine Guédon

Antoine Guédon

About

News

Publications

From Blobs to Spokes: High-Fidelity Surface Reconstruction via Oriented Gaussians

MAGICIAN: Efficient Long-Term Planning with Imagined Gaussians for Active Mapping

MILo: Mesh-In-the-Loop Gaussian Splatting for Detailed and Efficient Surface Reconstruction

MAtCha Gaussians: Atlas of Charts for High-Quality Geometry and Photorealism From Sparse Views

Gaussian Frosting: Editable Complex Radiance Fields with Real-Time Rendering

SuGaR: Surface-Aligned Gaussian Splatting for Efficient 3D Mesh Reconstruction and High-Quality Mesh Rendering

MACARONS: Mapping And Coverage Anticipation with RGB ONline Self-supervision

SCONE: Surface Coverage Optimization in uNknown Environments by Volumetric Integration