Research map / 3D vision / Face and shape
DiMeR: Disentangled Mesh Reconstruction Model with Normal-only Geometry Training
Lutao Jiang, Jiantao Lin, Kanghao Chen, Wenhang Ge, Xin Yang, Yifan Jiang, Yuanhuiyi Lyu, Xu Zheng, LI JING, Yinchuan Li, Ying-Cong Chen
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 |
|---|---|---|---|
| use only normal maps as geometry input (drop RGB) (ablation) | DiMeR · GSO | Chamfer Distance: 0.041 -> 0.028 | Improved |
| disentangle geometry from texture output (ablation) | DiMeR · GSO | Chamfer Distance: 0.043 -> 0.041 | Improved |
| add 3D regularization (eikonal + 3D GT loss) (ablation) | DiMeR · GSO | Chamfer Distance: 0.037 -> 0.028 | Improved |
| add PBR expectation loss (ablation) | DiMeR · GSO | Chamfer Distance: 0.039 -> 0.028 | Improved |
| remove deformation and weight MLP from FlexiCubes (ablation) | InstantMesh · GSO | Chamfer Distance: 0.045 -> 0.045; F1 0.964 -> 0.963; memory 73GB -> 48GB | No clear effect |
| raise SDF grid resolution from 128 to 192 (ablation) | DiMeR · GSO | Chamfer Distance: 0.032 -> 0.028 | Improved |
| add Gaussian noise to input normal maps (std 0.5) | DiMeR · GSO | Chamfer Distance: 0.028 -> 0.042 | Hurt |
| use StableNormal instead of GeoWizard as normal predictor | DiMeR · GSO | Chamfer Distance: 0.039 -> 0.032 | Improved |
| use joint normal from Zero123plus instead of StableNormal | DiMeR · GSO single-image-to-3D | Chamfer Distance: 0.052 -> 0.061 | Hurt |
| use ground-truth normals instead of predicted | DiMeR · GSO | Chamfer Distance: 0.032 -> 0.028 | Improved |
Closest papers by meaning
- MeshFormer : High-Quality Mesh Generation with 3D-Guided Reconstruction Model (NeurIPS 2024 · similarity 0.88)
- Mesh2Tex: Generating Mesh Textures from Image Queries (ICCV 2023 · similarity 0.88)
- Dens3R: A Foundation Model for 3D Geometry Prediction (ICLR 2026 · similarity 0.88)
- One-2-3-45: Any Single Image to 3D Mesh in 45 Seconds without Per-Shape Optimization (NeurIPS 2023 · similarity 0.88)
- GTR: Improving Large 3D Reconstruction Models through Geometry and Texture Refinement (ICLR 2025 · similarity 0.87)
- Fine-Detail Monocular Geometry Estimation with Self-Guided Sparse Volumetric Refinement (NeurIPS 2026 · 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.
