Research map / Point clouds / Point cloud classification
Space-Partitioning RANSAC
Daniel Barath, Gábor Valasek
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 space-partitioning verification | LO-RANSAC · Homography (Sacre Coeur), no SPRT | verification time (ms): 88.8 -> 40.0 | Improved |
| add space-partitioning verification | LO-RANSAC · Fundamental matrix, no SPRT | verification time (ms): 113.7 -> 50.3 | Improved |
| add space-partitioning verification | LO-RANSAC · Essential matrix, no SPRT | verification time (ms): 537.6 -> 251.2 | Improved |
| add space-partitioning verification on top of SPRT | LO-RANSAC with SPRT · Essential matrix | verification time (ms): 79.7 -> 29.4 | Improved |
| add space-partitioning verification on top of SPRT | LO-RANSAC with SPRT · Radial homography (Sun360) | average run-time (ms): 331.3 -> 134.7 | Improved |
| add space-partitioning verification on top of SPRT | LO-RANSAC with SPRT · Homography (Sacre Coeur) | average run-time (ms): 39.3 -> 31.2 | Improved |
| add space-partitioning verification | LO-RANSAC · Essential matrix | AUC@10: 0.65 -> 0.65 | No clear effect |
| add space-partitioning verification to SPRT | LO-RANSAC with SPRT · Essential matrix | AUC@10: 0.61 -> 0.63 | Mixed |
| raise early rejection threshold above 1 (1.6) for homography (ablation) | space-partitioning RANSAC · Homography | run-time: approximately 20% lower run-time | Improved |
| raise early rejection threshold to 1.2 for F and E (ablation) | space-partitioning RANSAC · Fundamental and essential matrix | processing time: 22% and 9% lower | Improved |
| use more grid cells (beyond 16) for F estimation (ablation) | space-partitioning RANSAC · Fundamental matrix | run-time: 16 cells fastest; time increases with cell number | Hurt |
| use 256 cells instead of other cell counts (ablation) | space-partitioning RANSAC · Homography | run-time: 50% of traditional run-time at 10000 iterations | Improved |
Closest papers by meaning
- Graph-Cut RANSAC: Local Optimization on Spatially Coherent Structures (TPAMI 2021 · similarity 0.85)
- Graph-Cut RANSAC (CVPR 2018 · similarity 0.83)
- Latent RANSAC (CVPR 2018 · similarity 0.83)
- Learning To Find Good Models in RANSAC (CVPR 2022 · similarity 0.83)
- PARSAC: Accelerating Robust Multi-Model Fitting with Parallel Sample Consensus (AAAI 2024 · similarity 0.81)
- RANSAC Scoring Done Right (NeurIPS 2026 · 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.
