Research map / 3D vision / Face and shape
MeshSDF: Differentiable Iso-Surface Extraction
Edoardo Remelli, Artem Lukoianov, Stephan Richter, Benoit Guillard, Timur Bagautdinov, Pierre Baque, Pascal Fua
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Add differentiable silhouette refinement of the latent code at inference (ablation) | MeshSDF (raw) · ShapeNet Core SVR | CD-l2 x10^3 (mean): 12.0 -> 7.8 | Improved |
| Add differentiable silhouette refinement of the latent code at inference (ablation) | MeshSDF (raw) · ShapeNet Core SVR | IoU (mean): 28 -> 32 | Improved |
| Add differentiable silhouette refinement of the latent code at inference (ablation) | MeshSDF (raw) · Pix3D Chairs SVR | IoU: 0.337 -> 0.407 | Improved |
| Add differentiable silhouette refinement of the latent code at inference (ablation) | MeshSDF (raw) · ShapeNet Core SVR | F-Score %: 91 -> 95 | Improved |
| Use MeshSDF latent parameterization instead of hand-crafted parameterizations for drag minimization | CFD shape optimization · CFD drag minimization, 8 cars | relative objective: 0.841 -> 0.675 vs PolyCube | Improved |
| Optimize vertices directly instead of through low-dimensional parameterization (ablation) | CFD shape optimization · CFD drag minimization | relative objective | Hurt |
| Fine-tune DeepSDF end-to-end on Chamfer distance via differentiable extraction | DeepSDF · chairs test | Chamfer distance: 0.00407 -> 0.00259 | Improved |
| Fine-tune DeepSDF end-to-end on Chamfer distance via differentiable extraction | DeepSDF · cars test | Chamfer distance: 0.00084 -> 0.00067 | Improved |
Later papers that compared with this method
1 later accepted paper reports MeshSDF as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- Sketch2Mesh: Reconstructing and Editing 3D Shapes From Sketches (ICCV 2021 · better on 1 of 1 shared result)
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- Pixel2Mesh — Pixel2Mesh: Generating 3D Mesh Models from Single RGB Images (ECCV 2018) · better on 4 of 4 shared results
- DeepSDF — DeepSDF: Learning Continuous Signed Distance Functions for Shape Representation (CVPR 2019) · better on 2 of 2 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- DeepMesh: Differentiable Iso-Surface Extraction (TPAMI 2024 · similarity 0.92)
- HSDF: Hybrid Sign and Distance Field for Modeling Surfaces with Arbitrary Topologies (NeurIPS 2022 · similarity 0.87)
- MIND: Material Interface Generation from UDFs for Non-Manifold Surface Reconstruction (NeurIPS 2025 · similarity 0.86)
- Iso-Points: Optimizing Neural Implicit Surfaces With Hybrid Representations (CVPR 2021 · similarity 0.86)
- DUDF: Differentiable Unsigned Distance Fields with Hyperbolic Scaling (CVPR 2024 · similarity 0.86)
- High Resolution UDF Meshing via Iterative Networks (NeurIPS 2025 · similarity 0.85)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
