AI Paper Atlas

Research map / Point clouds / Registration and matching

UMERegRobust – Universal Manifold Embedding Compatible Features for Robust Point Cloud Registration

Yuval Haitman, Amit Efraim, Joseph M Francos

ECCV 2024 · Point clouds › Registration and matching · approach: Learned feature-based registration · 4 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
replace Horn/RANSAC-style point-coordinate estimation with UME-based estimation and feature-correlation hypothesis selectionUMERegRobust · KITTI (strict precision)registration recall (%): 78.6 -> 87.8Improved
use UMERegRobust on large-rotation scans versus GCLUMERegRobust · RotKITTI (strict precision)registration recall (%): 28.8 -> 73.3Improved
use UMERegRobust versus GeoTransformer on large rotationsUMERegRobust · RotKITTI (strict precision)registration recall (%): 50.1 -> 73.3Improved
use UMERegRobust versus GCL on low-overlap large-translation scansUMERegRobust · LoKITTI (normal precision)registration recall (%): 72.3 -> 59.3Hurt
use UMERegRobust versus GCL on low-density LiDAR low overlapUMERegRobust · LonuScenes (strict precision)registration recall (%): 5.6 -> 56.3Improved
use UMERegRobust on indoor 3DMatch versus recent baselinesUMERegRobust · 3DMatchregistration recallNo clear effect
add UME-compatible losses to feature extractor training (ablation)UMERegRobust coloring module · KITTI and RotKITTIregistration recallImproved
add Sampling Equalizer Module to non-UME coloring (ablation)UMERegRobust coloring module · KITTI and RotKITTIregistration recallImproved
add Sampling Equalizer Module to GCL baseline (ablation)GCL · KITTI and RotKITTI (normal precision)registration recallMixed
match neighborhoods with UME descriptor and d_pF distance instead of point-wise l2 on UME point descriptors (ablation)local UME descriptors · keypoint matching PR curvematching precision: 30% deterioration with l2Improved
apply ICP refinement after registration (ablation)UMERegRobust · outdoor benchmarksregistration recall: 2% to 4% degradation without ICPImproved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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