Research map / 3D vision / Face and shape
NeuFrameQ: Neural Frame Fields for Scalable and Generalizable Anisotropic Quadrangulation
Ying-Tian Liu, Jiajun Li, Yu-Tao Liu, Xin Yu, Yuan-Chen Guo, Yan-Pei Cao, Ding Liang, Ariel Shamir, Song-Hai 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 |
|---|---|---|---|
| Use polyvector representation instead of raw frame vectors (ablation) | frame field regression · DFaust | mean angle error: 6.414 -> 6.214 | Improved |
| Use polyvector representation instead of raw frame vectors (ablation) | frame field regression · Our dataset | mean angle error: 11.916 -> 9.231 | Improved |
| Use polyvector representation instead of raw frame vectors (ablation) | frame field regression · DFaust | mean magnitude error: 0.0974 -> 0.0956 | Improved |
| Use polyvector representation instead of raw frame vectors (ablation) | frame field regression · Our dataset | mean magnitude error: 0.5627 -> 0.4247 | Improved |
| Add two-stage decomposition (direction regression then diffusion magnitude) (ablation) | polyvector regression · DFaust | mean angle error: 6.214 -> 5.316 | Improved |
| Add two-stage decomposition (direction regression then diffusion magnitude) (ablation) | polyvector regression · Our dataset | mean angle error: 9.231 -> 6.303 | Improved |
| Add two-stage decomposition (direction regression then diffusion magnitude) (ablation) | polyvector regression · DFaust | mean magnitude error: 0.0956 -> 0.0811 | Improved |
| Add two-stage decomposition (direction regression then diffusion magnitude) (ablation) | polyvector regression · Our dataset | mean magnitude error: 0.4247 -> 0.3289 | Improved |
| Use NeuFrameQ frame fields instead of LDF | frame field prediction · DFaust | mean angle error: 6.987 -> 5.316 | Improved |
| Use NeuFrameQ frame fields instead of IntrinsicVH | frame field prediction · DFaust | mean magnitude error: 0.1043 -> 0.0811 | Improved |
| Use NeuFrameQ quadrangulation instead of Instant-Meshes | quad remeshing · QuadWild-300 nicolo | chamfer distance (x1e-5): 0.269 -> 0.757 | Hurt |
| Use NeuFrameQ quadrangulation instead of QuadWild | quad remeshing · QuadWild-300 lock | chamfer distance (x1e-5): 0.587 -> 0.035 | Improved |
Closest papers by meaning
- PoNQ: a Neural QEM-based Mesh Representation (CVPR 2024 · similarity 0.86)
- Flatten Anything: Unsupervised Neural Surface Parameterization (NeurIPS 2024 · similarity 0.86)
- QUADify: Extracting Meshes with Pixel-level Details and Materials from Images (CVPR 2024 · similarity 0.86)
- MeshGPT: Generating Triangle Meshes with Decoder-Only Transformers (CVPR 2024 · similarity 0.86)
- MeshXL: Neural Coordinate Field for Generative 3D Foundation Models (NeurIPS 2024 · similarity 0.85)
- Mesh BDF: Barycentric Dominance Field for 3D Native Mesh Generation (NeurIPS 2026 · 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.
