AI Paper Atlas

Research map / Point clouds / Shape and surface completion

Learning and Matching Multi-View Descriptors for Registration of Point Clouds

Lei Zhou, Siyu Zhu, Zixin Luo, Tianwei Shen, Runze Zhang, Mingmin Zhen, Tian Fang, Long Quan

ECCV 2018 · Point clouds › Shape and surface completion · approach: Unsupervised shape correspondence · 50 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
use Fuseption-based FRN fusion instead of view pooling (ablation)MVDesc · HPatches patch verificationmAP: 0.883 -> 0.921Improved
use Fuseption-based FRN fusion instead of view pooling (ablation)MVDesc · HPatches patch retrievalmAP: 0.456 -> 0.530Improved
use Fuseption-based FRN fusion instead of view pooling (ablation)MVDesc · HPatches image matchingmAP: 0.312 -> 0.325Improved
use 3 views of a patch instead of a single viewMVDesc vs L2-Net · HPatches patch verificationmAP: 0.792 -> 0.921Improved
increase descriptor dimension (ablation)MVDesc · HPatches image matchingmAPMixed
use MVDesc-128 instead of MVDesc-32MVDesc · TUM descriptor matchingprecision: 0.664 -> 0.695Improved
use MVDesc-32 instead of CGFdescriptor matching · TUM descriptor matchingprecision: 0.447 -> 0.664Improved
add RMBP outlier rejection to CGF (ablation)CGF · ScanNet fragment registrationrecall: 0.350 -> 0.419Improved
add RMBP outlier rejection to 3DMatch (ablation)3DMatch · ScanNet fragment registrationrecall: 0.185 -> 0.145Hurt
add RMBP outlier rejection to MVDesc-128 (ablation)MVDesc-128 · ScanNet fragment registrationprecision: 0.910 -> 0.906Hurt
add RMBP outlier rejection to FPFH (ablation)FPFH · ScanNet fragment registrationrecall: 0.119 -> 0.272Improved
use RMBP instead of RANSAC, SMG or LPM at low inlier ratiooutlier rejection · TUM, ScanNet, EPFL point cloud pairsvalid registrationImproved

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