AI Paper Atlas

Research map / Learning theory and algorithms / Matrix and tensor methods

Generalized Shuffled Linear Regression

Feiran Li, Kent Fujiwara, Fumio Okura, Yasuyuki Matsushita

ICCV 2021 · Learning theory and algorithms › Matrix and tensor methods · approach: High-dimensional estimation theory · 8 citations (OpenAlex)

Read the paper

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 GSLR (outlier-rejecting one-to-one matching) instead of CPDCPD · Synthetic point clouds with outliersmean transformational error: 1.58e-2 -> 6.24e-3Improved
use GSLR instead of Trim-ICPTrim-ICP · Synthetic point clouds with outliersmean transformational error: 1.07 -> 6.24e-3Improved
use GSLR instead of ICPICP · Synthetic point clouds with partial overlapmean transformational error: 1.70e-1 -> 6.52e-2Improved
use GSLR instead of CPDCPD · Synthetic point clouds with partial overlapmedian transformational error: 9.09e-2 -> 3.78e-2Improved
use GSLR instead of FGRFGR · Synthetic point clouds with partial overlapmean transformational error: 1.07 -> 6.52e-2Improved
use GSLR instead of ICPICP · ETH Hauptgebaude real-world point cloudsmean transformational error: 5.82e-1 -> 1.40e-1Improved
use GSLR instead of CPDCPD · ETH Hauptgebaude real-world point cloudsmedian transformational error: 1.22e-1 -> 1.16e-1Improved
use GSLR instead of FGRFGR · ETH Hauptgebaude real-world point cloudsmean transformational error: 1.23 -> 1.40e-1Improved
fix inlier count k to all points (SLR) vs estimating k via Huber-skip (GSLR) (ablation)SLR · Image registration (SIFT points, homography)registration quality: SLR distorted on all trials -> GSLR matches ground truth visuallyImproved
use GSLR instead of ICP/Trim-ICP under coarse initializationICP and Trim-ICP · Bunny point clouds with 1/3 outlierssensitivity to initialization: fail drastically when coarsened -> stable under large initialization errorsImproved
use GSLR instead of ZoomOut-100 for shape matchingZoomOut-100 · FAUST/TOSCA shape matchinggeodesic error: similar accuracy; ZoomOut doubles eigenfunction dimensionsNo clear effect
use GSLR instead of ICP, BCICP, ZoomOut for shape matchingICP, BCICP, ZoomOut · FAUST shape matchingcoverage: covering more than 80% vertices on all the pairsImproved

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