Research map / Learning theory and algorithms / Combinatorial and search optimization
Efficient Subpixel Refinement With Symbolic Linear Predictors
Vincent Lui, Jonathon Geeves, Winston Yii, Tom Drummond
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 symbolic representations to learn the linear predictor | Linear Predictors for synthetic subpixel refinement · Synthetic experiment, 9×9 patch, m=5000 | RMSE | Improved |
| Use symbolic representations instead of conventional LP training | Linear Predictors for subpixel refinement · Synthetic experiment | training time | Improved |
| Use the symbolic predictor for refinement | Subpixel refinement · Localization experiment | ATE (mm): 8.1 -> 2.3 | Improved |
| Select the best 250 keypoints by predicted error rather than using the worst 250 | Symbolic Linear Predictors · Localization experiment | ATE (mm): 3.1 -> 2.3 | Improved |
| Combine symbolic predictors with DCT-25 (ablation) | Symbolic Linear Predictors · Hannover image pairs | RMSE | Mixed |
| Retain fewer, low-frequency DCT coefficients (ablation) | Linear Predictors · Hannover image pairs | RMSE | Improved |
| Increase the number of sample warps (ablation) | Linear Predictors · Synthetic experiment | RMSE | Improved |
| Increase the patch size (ablation) | Linear Predictors · Synthetic experiment | RMSE | Improved |
| Increase maximum parameter variation during training (ablation) | Linear Predictors · Synthetic experiment | RMSE | Hurt |
| Use symbolic learning instead of conventional learning for the same LP (ablation) | Linear Predictors · Synthetic experiment | RMSE | No clear effect |
| Use DCT-25 with symbolic learning instead of DCT-25 alone (ablation) | Linear Predictors · Synthetic experiment | RMSE | No clear effect |
Closest papers by meaning
- ALRe: Outlier Detection for Guided Refinement (ECCV 2020 · similarity 0.84)
- Learning Invariant Inter-pixel Correlations for Superpixel Generation (AAAI 2024 · similarity 0.84)
- FastJAM: a Fast Joint Alignment Model for Images (NeurIPS 2025 · similarity 0.82)
- UniDemoiré: Towards Universal Image Demoiréing with Data Generation and Synthesis (AAAI 2025 · similarity 0.82)
- AUGCAL: Improving Sim2Real Adaptation by Uncertainty Calibration on Augmented Synthetic Images (ICLR 2024 · similarity 0.81)
- Optimize & Reduce: A Top-Down Approach for Image Vectorization (AAAI 2024 · similarity 0.81)
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.
