Research map / Learning theory and algorithms / Kernels and regression theory
A Generalization of Otsu’s Method and Minimum Error Thresholding
Jonathan T. Barron
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 |
|---|---|---|---|
| add the weighted-percentile prior (kappa, omega) to GHT (ablation) | GHT · H-DIBCO 2016 | F1 x100: 87.16 -> 88.77 | Improved |
| add the weighted-percentile prior (kappa, omega) to GHT (ablation) | GHT · H-DIBCO 2016 | DRD: 5.04 -> 3.99 | Improved |
| add the Gaussian mixture components and priors to the weighted percentile only model (ablation) | GHT · H-DIBCO 2016 | F1 x100: 76.77 -> 88.77 | Improved |
| generalize MET with the variance prior (nu, tau) and percentile prior (ablation) | Minimum Error Thresholding · H-DIBCO 2016 | F1 x100: 60.40 -> 88.77 | Improved |
| generalize Otsu's method with the percentile prior and tuned hyperparameters (ablation) | Otsu's method · H-DIBCO 2016 | F1 x100: 87.19 -> 88.77 | Improved |
| generalize Otsu's method with the percentile prior and tuned hyperparameters (ablation) | Otsu's method · H-DIBCO 2016 | DRD: 5.04 -> 3.99 | Improved |
| use the global-threshold GHT instead of the best neural per-pixel entry (Hassaine et al. - 3) | GHT · H-DIBCO 2016 | F1 x100: 88.72 -> 88.77 | No clear effect |
| use GHT instead of the best neural per-pixel entry (Hassaine et al. - 3) | GHT · H-DIBCO 2016 | DRD: 3.86 -> 3.99 | Hurt |
| use GHT instead of Otsu's method as published | GHT · H-DIBCO 2016 | PSNR: 17.80 -> 18.55 | Improved |
| choose threshold per image by oracle instead of one tuned global GHT setting | GHT · H-DIBCO 2016 | F1 x100: 88.77 -> 90.69 | Improved |
Closest papers by meaning
- GBHT: Gradient Boosting Histogram Transform for Density Estimation (ICML 2021 · similarity 0.84)
- Distribution-Free Calibration Guarantees for Histogram Binning without Sample Splitting (ICML 2021 · similarity 0.83)
- Logarithmic Lenses: Exploring Log RGB Data for Image Classification (CVPR 2024 · similarity 0.82)
- Hierarchical Histogram Threshold Segmentation - Auto-terminating High-detail Oversegmentation (CVPR 2024 · similarity 0.82)
- Image Synthesis Under Limited Data: A Survey and Taxonomy (IJCV 2025 · similarity 0.82)
- A Surprisingly Simple Approach to Generalized Few-Shot Semantic Segmentation (NeurIPS 2024 · similarity 0.82)
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.
