Research map / Point clouds / Shape and surface completion
Unbiased Gradient Estimation for Differentiable Surface Splatting via Poisson Sampling
Jan U. Müller, Michael Weinmann, Reinhard Klein
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 |
|---|---|---|---|
| increase samples per pixel from 1 to 40 (ablation) | Ours · Bunny | Hausdorff distance: 8.287 -> 0.442 | Improved |
| increase samples per pixel from 1 to 40 (ablation) | Ours · Bunny | Chamfer distance: 6.766 -> 0.125 | Improved |
| increase samples per pixel from 20 to 40 (ablation) | Ours · Bunny | Chamfer distance: 0.105 -> 0.125 | Hurt |
| increase samples per pixel from 20 to 40 (ablation) | Ours · Bunny | Hausdorff distance: 1.07 -> 0.442 | Improved |
| increase number of views from 16 to 64 (ablation) | Ours · Bunny | Hausdorff distance: 49.159 -> 1.097 | Improved |
| increase number of views from 124 to 300 (ablation) | Ours · Bunny | Hausdorff distance: 0.442 -> 0.409 | No clear effect |
| use unbiased Poisson-sampling gradient instead of DSS surrogate gradient | DSS · Bunny shape reconstruction | Hausdorff distance: 7.73 -> 0.442 | Improved |
| use unbiased Poisson-sampling gradient instead of Pulsar | Pulsar · Yoga1 shape reconstruction | Hausdorff distance: 25241.71 -> 3.355 | Improved |
| use unbiased Poisson-sampling gradient instead of SynSin | SynSin · Teapot shape reconstruction | Chamfer distance: 4.771 -> 0.289 | Improved |
| use unbiased Poisson-sampling gradient instead of Pulsar | Pulsar · Teapot shape reconstruction | Hausdorff distance: 17.036 -> 0.453 | Improved |
| refine SLAM reconstruction with the splatting renderer | SLAM output · room-scale scene (Bode et al.) | PSNR: 18.095 -> 23.257 | Improved |
| refine SLAM reconstruction with the splatting renderer | SLAM output · room-scale scene (Bode et al.) | SSIM: 0.681 -> 0.799 | Improved |
Closest papers by meaning
- SplatFormer: Point Transformer for Robust 3D Gaussian Splatting (ICLR 2025 · similarity 0.87)
- UniPre3D: Unified Pre-training of 3D Point Cloud Models with Cross-Modal Gaussian Splatting (CVPR 2025 · similarity 0.87)
- UniGS: Unified Geometry-Aware Gaussian Splatting for Multimodal Rendering (NeurIPS 2026 · similarity 0.87)
- Evolving High-Quality Rendering and Reconstruction in a Unified Framework with Contribution-Adaptive Regularization (CVPR 2025 · similarity 0.87)
- ComPC: Completing a 3D Point Cloud with 2D Diffusion Priors (ICLR 2025 · similarity 0.86)
- 3DMambaIPF: A State Space Model for Iterative Point Cloud Filtering via Differentiable Rendering (AAAI 2025 · 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.
