Research map / Learning theory and algorithms / Kernels and regression theory
Incomplete Gamma Kernels: Generalizing Locally Optimal Projection Operators
Patrick Stotko, 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 |
|---|---|---|---|
| replace Gaussian localization kernel with actual K_LOP in WLOP density weights (simple scheme) | WLOP · Bird irregular target | point cloud regularity sigma_Q: better than WLOP in 76.7% of combinations | Improved |
| solve full linear system for density weights instead of reciprocal density | WLOP simple scheme · Bird irregular target | point cloud regularity sigma_Q: outperforms WLOP in 99.3% and simple scheme in 98.7% | Improved |
| use fitted kernel approximation parameters instead of CLOP parameters | CLOP kernel approximation · K_LOP std ratio d=1 | std ratio: 0.7931 -> 0.9917 | Improved |
| use consistent approximation constraint | kernel approximation · K_LOP std ratio d=4 | std ratio: 0.9640 -> 1.0440; unbiased at d to infinity (0.8881 -> 1) | Mixed |
| use our kernel approximation instead of CLOP in WLOP smoothing | WLOP smoothing · Block model | point-surface error at edges | Improved |
| apply optimal L1 scaling correction to kernel approximation | kernel approximation · K_LOP L1 distance | L1 kernel error | Improved |
| replace L2 loss with localized LOP/Gaussian loss in normal filtering | normal filtering · Gargoyle and Box | feature preservation | Improved |
| replace Gaussian prior kernel with K_LOP | TotalDenoising · ModelNet40 sigma=1.5% | Chamfer distance (x1e-4): 0.466 -> 0.424 | Improved |
| replace Gaussian prior kernel with K_LOP | TotalDenoising · ModelNet40 sigma=1% | Chamfer distance (x1e-4): 0.343 -> 0.282 | Improved |
| replace Gaussian prior kernel with K_LOP | TotalDenoising · ModelNet40 sigma=0.5% | Chamfer distance (x1e-4): 0.173 -> 0.175 | No clear effect |
| replace Gaussian prior kernel with K_LOP | TotalDenoising · Paris-rue-Madame | oversmoothing and shrinkage | Improved |
Closest papers by meaning
- Kernel Bayesian Inference with Posterior Regularization (NeurIPS 2016 · similarity 0.85)
- Optimal Rates of Kernel Ridge Regression under Source Condition in Large Dimensions (JMLR 2025 · similarity 0.85)
- On the Posterior Distribution in Denoising: Application to Uncertainty Quantification (ICLR 2024 · similarity 0.84)
- Generalization error of spectral algorithms (ICLR 2024 · similarity 0.84)
- Kernel Regression in Structured Non-IID Settings: Theory and Implications for Denoising Score Learning (NeurIPS 2025 · similarity 0.84)
- Generalization in Kernel Regression Under Realistic Assumptions (ICML 2024 · similarity 0.84)
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.
