Research map / 3D vision / NeRF and radiance fields
Mesh2NeRF: Direct Mesh Supervision for Neural Radiance Field Representation and Generation
Yujin Chen, Yinyu Nie, Benjamin Ummenhofer, Reiner Birkl, Michael Paulitsch, Matthias Müller, Matthias Niessner
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 |
|---|---|---|---|
| supervise density and color per ray sample directly from the mesh instead of pixel rendering loss (Instant NGP) | Instant NGP · ABO single scene fitting | PSNR: 30.16 -> 33.28 | Improved |
| direct mesh supervision applied to vanilla NeRF | NeRF · ABO single scene fitting | PSNR: 25.09 -> 32.40 | Improved |
| direct mesh supervision applied to NeRF on Poly Haven | NeRF · Poly Haven | PSNR: 21.24 -> 22.75 | Improved |
| direct mesh supervision applied to TensoRF on Poly Haven | TensoRF · Poly Haven | PSNR: 23.32 -> 23.97 | Improved |
| replace rendering loss with Mesh2NeRF supervision in SSDNeRF (1-view cars) | SSDNeRF · ShapeNet Cars 1-view | PSNR: 21.09 -> 21.78 | Improved |
| replace rendering loss with Mesh2NeRF supervision in SSDNeRF (4-view cars) | SSDNeRF · ShapeNet Cars 4-view | PSNR: 26.54 -> 26.51 | No clear effect |
| replace rendering loss with Mesh2NeRF supervision in SSDNeRF (2-view chairs) | SSDNeRF · ShapeNet Chairs 2-view | PSNR: 19.65 -> 22.22 | Improved |
| add integral loss with weight 10 (ablation) | Mesh2NeRF NGP · Poly Haven | PSNR: 21.93 -> 25.30 | Improved |
| use integral loss weight 100 instead of 10 (ablation) | Mesh2NeRF NGP · Poly Haven | PSNR: 25.30 -> 13.70 | Hurt |
| use only integral loss without alpha and color loss (ablation) | Mesh2NeRF NGP · Poly Haven | PSNR: 25.30 -> 12.90 | Hurt |
| reduce surface thickness from 0.0200 to 0.0050 (ablation) | Mesh2NeRF NeRF · ABO | PSNR: 27.18 -> 28.40 | Improved |
| reduce surface thickness from 0.0050 to 0.0025 (ablation) | Mesh2NeRF NeRF · ABO | LPIPS: PSNR 28.40 -> 28.93; LPIPS 0.049 -> 0.056 | Mixed |
Closest papers by meaning
- Learning Neural Duplex Radiance Fields for Real-Time View Synthesis (CVPR 2023 · similarity 0.91)
- GRF: Learning a General Radiance Field for 3D Representation and Rendering (ICCV 2021 · similarity 0.90)
- GRAF: Generative Radiance Fields for 3D-Aware Image Synthesis (NeurIPS 2020 · similarity 0.90)
- VD-NeRF: Visibility-Aware Decoupled Neural Radiance Fields for View-Consistent Editing and High-Frequency Relighting (TPAMI 2025 · similarity 0.89)
- SIGNeRF: Scene Integrated Generation for Neural Radiance Fields (CVPR 2024 · similarity 0.89)
- HyRF: Hybrid Radiance Fields for Memory-efficient and High-quality Novel View Synthesis (NeurIPS 2025 · similarity 0.89)
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.
