Research map / 3D vision / Gaussian splatting
NeuMesh++: Toward Versatile and Efficient Volumetric Editing With Disentangled Neural Mesh-Based Implicit Field
Chong Bao, Yuan Li, Bangbang Yang, Yujun Shen, Hujun Bao, Zhaopeng Cui, Yinda Zhang, Guofeng Zhang
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 |
|---|---|---|---|
| interpolate features after MLP decoding (post-interpolation) instead of before (ablation) | NeuMesh++ · NeRF Synthetic 360 | PSNR: 32.17 -> 32.32 | Improved |
| interpolate after MLP decoding instead of before (ablation) | NeuMesh++ · DTU | PSNR: 28.37 -> 28.39 | Improved |
| add normal conditioning (ablation) | NeuMesh++ · NeRF Synthetic 360 | PSNR: 32.32 -> 32.32 | No clear effect |
| add normal conditioning (ablation) | NeuMesh++ · texture swapping (qualitative) | view-dependent shiny effects | Improved |
| add per-vertex modification color (ablation) | NeuMesh++ · texture painting (qualitative) | painting noise | Improved |
| add distillation from teacher during training (ablation) | NeuMesh++ · NeRF Synthetic 360 | PSNR: 32.324 -> 32.2740 | No clear effect |
| use full-size neural codes instead of halved codes (LITE) (ablation) | NeuMesh++ · NeRF Synthetic 360 | PSNR: 32.06 -> 32.32 | Improved |
| raise mesh resolution from 100 to 200 (ablation) | NeuMesh++ · NeRF Synthetic 360 | PSNR: 30.352 -> 31.900 | Improved |
| raise mesh resolution from 200 to 300 (ablation) | NeuMesh++ · NeRF Synthetic 360 | PSNR: 31.900 -> 32.324 | Improved |
| raise KNN neighbors beyond 4 (ablation) | NeuMesh++ · NeRF Synthetic 360 | PSNR | No clear effect |
| replace NeuMesh with NeuMesh++ | NeuMesh · NeRF Synthetic 360 | PSNR: 26.21 -> 32.32 | Improved |
| replace NeuMesh with NeuMesh++ | NeuMesh · texture painting time | seconds: 5492 -> 1 | Improved |
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.
- NeuMesh — NeuMesh: Learning Disentangled Neural Mesh-Based Implicit Field for Geometry and Texture Editing (ECCV 2022) · better on 4 of 4 shared results
- NeuS — NeuS: Learning Neural Implicit Surfaces by Volume Rendering for Multi-view Reconstruction (NeurIPS 2021) · better on 1 of 1 shared result
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
- NeuMesh: Learning Disentangled Neural Mesh-Based Implicit Field for Geometry and Texture Editing (ECCV 2022 · similarity 0.97)
- SINE: Semantic-Driven Image-Based NeRF Editing With Prior-Guided Editing Field (CVPR 2023 · similarity 0.90)
- NeuTex: Neural Texture Mapping for Volumetric Neural Rendering (CVPR 2021 · similarity 0.89)
- NeRF-Editing: Geometry Editing of Neural Radiance Fields (CVPR 2022 · similarity 0.89)
- LAENeRF: Local Appearance Editing for Neural Radiance Fields (CVPR 2024 · similarity 0.87)
- Feature 3DGS: Supercharging 3D Gaussian Splatting to Enable Distilled Feature Fields (CVPR 2024 · similarity 0.86)
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.
