Research map / 3D vision / Face and shape
Learning Unsigned Distance Functions from Multi-view Images with Volume Rendering Priors
Wenyuan Zhang, Kanle Shi, Yu-Shen Liu, Zhizhong Han
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 |
|---|---|---|---|
| replace handcrafted renderer equations with a learned volume rendering prior | UDF rendering · ShapeNet GT UDF rendering | Depth-L1: 1.28 -> 0.41 | Improved |
| learned prior versus handcrafted renderers for UDF inference from RGB | NeuralUDF/NeUDF · DeepFashion3D | Chamfer Distance: 2.01 -> 1.71 | Improved |
| use sampling interval instead of coordinates as prior input (ablation) | volume rendering prior · DeepFashion3D | CD-L1: 2.92 -> 1.71 | Improved |
| include sampling interval as input alongside unsigned distances (ablation) | volume rendering prior · DeepFashion3D | CD-L1: 1.87 -> 1.71 | Improved |
| train the prior with RGBD supervision instead of depth only (ablation) | volume rendering prior · DeepFashion3D | CD-L1: 1.71 -> 2.02 | Hurt |
| train the prior with RGB supervision only (ablation) | volume rendering prior · DeepFashion3D | CD-L1: Divergent | Hurt |
| fine-tune the prior with RGB images at test time (ablation) | volume rendering prior · DeepFashion3D | CD-L1: Divergent | Hurt |
| use more than two training shapes for the prior (ablation) | volume rendering prior · DeepFashion3D and DTU | Chamfer Distance: 1.71 -> 1.72 (100 shapes) | No clear effect |
| different choices of training shape pair (ablation) | volume rendering prior · DeepFashion3D | Chamfer Distance: 1.71 to 1.78 across sets | No clear effect |
| neighboring window size 30 versus very small (1) or large (50) (ablation) | volume rendering prior · DeepFashion3D example shape | Chamfer Distance: 2.11 -> 1.59 | Improved |
| learn the prior for SDF instead of UDF | volume rendering prior · DeepFashion3D SDF example | Chamfer Distance: 0.73 -> 0.65 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 3 of 3 on most shared results.
- NeuS — NeuS: Learning Neural Implicit Surfaces by Volume Rendering for Multi-view Reconstruction (NeurIPS 2021) · better on 5 of 5 shared results
- NeuralUDF — NeuralUDF: Learning Unsigned Distance Fields for Multi-View Reconstruction of Surfaces With Arbitrary Topologies (CVPR 2023) · better on 4 of 4 shared results
- NeUDF — NeUDF: Leaning Neural Unsigned Distance Fields With Volume Rendering (CVPR 2023) · better on 3 of 3 shared results
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
- 2S-UDF: A Novel Two-stage UDF Learning Method for Robust Non-watertight Model Reconstruction from Multi-view Images (CVPR 2024 · similarity 0.91)
- DUDF: Differentiable Unsigned Distance Fields with Hyperbolic Scaling (CVPR 2024 · similarity 0.90)
- Sensing Surface Patches in Volume Rendering for Inferring Signed Distance Functions (AAAI 2025 · similarity 0.90)
- GaussianUDF: Inferring Unsigned Distance Functions through 3D Gaussian Splatting (CVPR 2025 · similarity 0.89)
- High Resolution UDF Meshing via Iterative Networks (NeurIPS 2025 · similarity 0.89)
- Neural Signed Distance Function Inference through Splatting 3D Gaussians Pulled on Zero-Level Set (NeurIPS 2024 · 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.
