Lightnews — Scholar-powered news

Reposted by Antoine Guédon

Imagine-ENPC @imagineenpc.bsky.social · 6d

Familiar names among #ICCV2025 Outstanding Reviewers from our team 😇
Antoine Guédon @antoine-guedon.bsky.social
Sinisa Stekovic
Renaud Marlet
👏
@iccv.bsky.social
iccv.thecvf.com/Conferences/...

2025 ICCV Program Committee

iccv.thecvf.com

5 21

Antoine Guédon @antoine-guedon.bsky.social · Sep 8

11/11 📚 Resources:
📄 Paper: arxiv.org/abs/2506.24096
💻 Code: github.com/Anttwo/MILo
🌐 Project Page: anttwo.github.io/milo/

Huge thanks to my amazing co-authors and the supporting institutions! 🙏

MILo

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

anttwo.github.io

1

Antoine Guédon @antoine-guedon.bsky.social · Sep 8

10/n📺Video:
See MILo in action!
Our presentation video showcases the differentiable pipeline and reconstruction results across various scenes.

🔗 YouTube video: www.youtube.com/watch?v=rOBs...

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

YouTube video by Antoine Guédon

www.youtube.com

1 1

Antoine Guédon @antoine-guedon.bsky.social · Sep 8

9/n🎮Interactive Gallery:
Check out our interactive, online 3D viewer with both mesh and Gaussian representations!

🔗 Gallery: anttwo.github.io/milo/#galler...

MILo

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

anttwo.github.io

1

Antoine Guédon @antoine-guedon.bsky.social · Sep 8

8/n📈Optional depth-order regularization:
For even cleaner backgrounds, we propose an optional loss using DepthAnythingV2 that enforces depth ordering consistency.

This drastically improves background geometry quality!

1

Antoine Guédon @antoine-guedon.bsky.social · Sep 8

7/n🎨Animation & Editing:
Since Gaussians align with the extracted mesh surface, any mesh modification can easily be propagated to the Gaussians!

We include in the code a Blender addon for easy editing and animation - no coding required.

1

Antoine Guédon @antoine-guedon.bsky.social · Sep 8

6/n🔧Plug-and-play design:
MILo can be integrated into any Gaussian Splatting pipeline!

We provide simple differentiable functions that take Gaussian parameters as input and return meshes.

Perfect for adding differentiable surface processing to your 3DGS projects!

1

Antoine Guédon @antoine-guedon.bsky.social · Sep 8

5/n🎯Scalability advantage:
MILo reconstructs full scenes including all background elements, not just foregrounds.

To achieve this efficiency, we select only surface-likely Gaussians by repurposing the importance sampling from Mini-Splatting2.

1 1

Antoine Guédon @antoine-guedon.bsky.social · Sep 8

4/n📊Results:
✅ Higher quality meshes with significantly fewer vertices
✅ 60-350MB mesh sizes (vs GBs in other methods)
✅ Complete scene reconstruction (including backgrounds)
✅ Better performance on benchmarks

Efficiency meets quality!

1

Antoine Guédon @antoine-guedon.bsky.social · Sep 8

3/n🏗️How MILo works:
1️⃣ Each Gaussian spawns pivots
2️⃣ Delaunay triangulation connects pivots
3️⃣ SDF values assigned to pivots
4️⃣ Differentiable Marching Tetrahedra extracts mesh

The pipeline is differentiable, enabling mesh supervision to improve Gaussian configurations!

1

Antoine Guédon @antoine-guedon.bsky.social · Sep 8

2/n🔗Key innovation: differentiable mesh extraction at every training iteration

Unlike previous methods, MILo extracts vertex locations and connectivity purely from Gaussian parameters, allowing gradient flow from mesh back to Gaussians. This creates a powerful feedback loop!

1 1

Antoine Guédon @antoine-guedon.bsky.social · Sep 8

1/n🚀Gaussians > Differentiable function > Mesh?
Check out our new work: MILo: Mesh-In-the-Loop Gaussian Splatting!

🎉Accepted to SIGGRAPH Asia 2025 (TOG)
MILo is a novel differentiable framework that extracts meshes directly from Gaussian parameters during training.

🧵👇

3 7 20

Antoine Guédon @antoine-guedon.bsky.social · Jun 13

I'm at #CVPR2025 to present our paper 🍵MAtCha Gaussians🍵, today Friday afternoon, Hall D, Poster 53!

If you're in Nashville and want to discuss detailed 3D mesh reconstruction from sparse or dense RGB images, let's connect!

@kyotovision.bsky.social

Antoine Guédon @antoine-guedon.bsky.social · Apr 3

💻We've released the code for our #CVPR2025 paper MAtCha!

🍵MAtCha reconstructs sharp, accurate and scalable meshes of both foreground AND background from just a few unposed images (eg 3 to 10 images)...

...While also working with dense-view datasets (hundreds of images)!

3 8

Reposted by Antoine Guédon

#CVPR2026 @cvprconference.bsky.social · May 10

Behind every great conference is a team of dedicated reviewers. Congratulations to this year’s #CVPR2025 Outstanding Reviewers!

cvpr.thecvf.com/Conferences/...

12 48

Reposted by Antoine Guédon

Imagine-ENPC @imagineenpc.bsky.social · Apr 30

#CVPR2025 Fri June 13 (PM) ✨ Highlight
🍵 MAtCha Gaussians: Atlas of Charts for High-Quality Geometry and Photorealism From Sparse Views
@antoine-guedon.bsky.social @kyotovision.bsky.social
📄 pdf: arxiv.org/abs/2412.06767
🌐 webpage: anttwo.github.io/matcha/

1 2 9

Antoine Guédon @antoine-guedon.bsky.social · Apr 3

I actually saw him dancing on a bench 😱
anttwo.github.io/frosting/

2

Antoine Guédon @antoine-guedon.bsky.social · Apr 3

And the fact that this pipeline makes it possible to get sharp meshes from sparse unposed imgs means 2 things:

1. MASt3R-SfM is so good, it's crazy... I love it.

2. The regularization we introduce seems to really help the representation to stabilize, even though the constraints are very sparse

1

Antoine Guédon @antoine-guedon.bsky.social · Apr 3

You're entirely right!
And actually MASt3R-SfM does the tougher part of the job, clearly 😁
I just meant that both can be used in a unified pipeline for getting sharp meshes from unposed images.

1 2

Antoine Guédon @antoine-guedon.bsky.social · Apr 3

This work was done in collaboration with Tomoki Ichikawa, Kohei Yamashita and Professor Ko Nishino from @kyotovision.bsky.social well as @imagineenpc.bsky.social

🌐Website: anttwo.github.io/matcha/

💻Code: github.com/Anttwo/MAtCha

MAtCha

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

anttwo.github.io

Antoine Guédon @antoine-guedon.bsky.social · Apr 3

🔑 Key point #3: We also introduce a novel “depth-order” regularization that leverages depth maps estimated with a monodepth estimator.

The depth maps can be multi-view inconsistent, no problem!

MAtCha still gets smooth, detailed background while preserving foreground details.

1 1

Antoine Guédon @antoine-guedon.bsky.social · Apr 3

🔑 Key point #2: Inspired by Gaussian Opacity Fields, we developed a new mesh extraction method for 2DGS.

It properly handles both foreground and background geometry while being lightweight if needed (only 150-350MB).

No post-processing mesh decimation is required!

1

Antoine Guédon @antoine-guedon.bsky.social · Apr 3

🔑 Key point #1: Our novel optimization pipeline is robust to sparse-view inputs (as few as 3 to 10 images) but also scales to dense-view scenarios (hundreds of views).

No more choosing between sparse or dense methods!

1

Antoine Guédon @antoine-guedon.bsky.social · Apr 3

MAtCha introduces a novel surface representation that reconstructs high-quality 3D meshes with photorealistic rendering from just a handful of images.

💡Our key idea: model scene geometry as an Atlas of Charts and refine it with 2D Gaussian surfels.

1

Antoine Guédon @antoine-guedon.bsky.social · Apr 3

💻We've released the code for our #CVPR2025 paper MAtCha!

🍵MAtCha reconstructs sharp, accurate and scalable meshes of both foreground AND background from just a few unposed images (eg 3 to 10 images)...

...While also working with dense-view datasets (hundreds of images)!

4 16 40

Reposted by Antoine Guédon

David Picard @davidpicard.bsky.social · Mar 21

🔥🔥🔥 CV Folks, I have some news! We're organizing a 1-day meeting in center Paris on June 6th before CVPR called CVPR@Paris (similar as NeurIPS@Paris) 🥐🍾🥖🍷

Registration is open (it's free) with priority given to authors of accepted papers: cvprinparis.github.io/CVPR2025InPa...

Big 🧵👇 with details!

8 51 140