Research map / 3D vision / Face and shape
NeuralSlice: Neural 3D Triangle Mesh Reconstruction via Slicing 4D Tetrahedral Meshes
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Represent 3D shape as slice of deformable 4D tetrahedral mesh (NEURALSLICE) vs explicit template baselines | single-template explicit reconstruction · ShapeNet 13 categories | Chamfer distance ×10^3: 0.537 vs best explicit baseline MNA-25 0.671 | Helped |
| Add Laplacian regularization (Llap) to output mesh (ablation) | NEURALSLICE · ShapeNet | normal consistency: NC 0.193 without vs 0.602 full | Helped |
| Add Laplacian regularization (Llap) (ablation) | NEURALSLICE · ShapeNet | Chamfer distance ×10^3: 0.449 without vs 0.538 full | Hurt |
| Add Lipschitz normalization (LipNorm) (ablation) | NEURALSLICE · ShapeNet | Chamfer distance ×10^3: 0.784 without vs 0.538 full | Helped |
| Add Lipschitz normalization (LipNorm) (ablation) | NEURALSLICE · ShapeNet | normal consistency: NC 0.451 without vs 0.602 full | Helped |
| Add subdivision-based refinement network (ablation) | NEURALSLICE · ShapeNet | Chamfer distance ×10^3: 1.819 without vs 0.538 full | Helped |
| Add subdivision-based refinement network (ablation) | NEURALSLICE · ShapeNet | normal consistency: NC 0.206 without vs 0.602 full | Helped |
| Use 4D template T2 instead of T1 or T3 (ablation) | NEURALSLICE · ShapeNet | Chamfer distance ×10^3: 0.538 (T2) vs 0.578 (T1) and 0.593 (T3) | Helped |
| Use T1 template instead of T2 (ablation) | NEURALSLICE · ShapeNet | normal consistency: NC 0.723 (T1) vs 0.602 (T2) | Helped |
| NEURALSLICE vs implicit methods (ConvOcc, Occ) under noisy input | explicit slicing representation · ShapeNet noisy input | inference time (s): 0.025 vs 0.261 (ConvOcc), 0.316 (Occ) | Helped |
| NEURALSLICE vs implicit methods (ConvOcc, Occ) under noisy input | explicit slicing representation · ShapeNet noisy input | mean Chamfer distance ×10^3: 0.538 vs 0.891 (ConvOcc), 1.630 (Occ) | Helped |
Closest papers by meaning
- Neural Template: Topology-Aware Reconstruction and Disentangled Generation of 3D Meshes (CVPR 2022 · similarity 0.89)
- Learning Deformable Tetrahedral Meshes for 3D Reconstruction (NeurIPS 2020 · similarity 0.89)
- Neural Parts: Learning Expressive 3D Shape Abstractions With Invertible Neural Networks (CVPR 2021 · similarity 0.88)
- MeshGPT: Generating Triangle Meshes with Decoder-Only Transformers (CVPR 2024 · similarity 0.87)
- Common3D: Self-Supervised Learning of 3D Morphable Models for Common Objects in Neural Feature Space (CVPR 2025 · similarity 0.86)
- PoNQ: a Neural QEM-based Mesh Representation (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.