AI Paper Atlas

Research map / Label-efficient and robust learning / Recognition and long-tail

Multi-Class Model Fitting by Energy Minimization and Mode-Seeking

Daniel Barath, Jiri Matas

ECCV 2018 · Label-efficient and robust learning › Recognition and long-tail · approach: Confidence calibration objectives · 49 citations (OpenAlex)

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 changedWhereMeasuredEffect
add the mode-seeking label replacement move to PEARL (ablation)PEARL · Synthetic 3 lines + 200 outliersprobability of returning correct instance count: 0.1 -> 0.6Improved
add the mode-seeking label replacement move to PEARL (ablation)PEARL · Synthetic 3 lines + 200 outliersprocessing time and convergence energyImproved
add mode-seeking, which suppresses similar instances (ablation)PEARL with label cost only · Synthetic 3 lines + 200 outliers (noise sigma varied)robustness to noiseImproved
use automatic fixed parameters instead of per-test tuned parameters (ablation)Multi-X · AdelaideRMF homographyaverage misclassification error (%): 1.37 -> 9.72Hurt
use Multi-X instead of T-Linkage for 1000-point line and circle fittingT-Linkage · Line and circle fittingprocessing time (s): 102.8 -> 5.1Improved
use Multi-X instead of PEARL for plane and cylinder fitting on case (6)PEARL · Plane and cylinder fitting (LIDAR)misclassification error (%): 17.35 -> 21.72Hurt
use Multi-X instead of T-Linkage on two-view motion cases (7)T-Linkage · AdelaideRMF motion (tuned per pair)average misclassification error (%): 1.06 -> 5.32Hurt

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy