AI Paper Atlas

Research map / 3D vision / Face and shape

Unsupervised Deep Learning for Structured Shape Matching

Jean-Michel Roufosse, Abhishek Sharma, Maks Ovsjanikov

ICCV 2019 · 3D vision › Face and shape · approach: Primitive-based shape abstraction · 161 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
combine all four penalties instead of E3 alone (ablation)SURFMNet · FAUSTmean geodesic error: 0.073 -> 0.020Improved
add E1 and E2 to E3 and E4 (vs E1,E3,E4 etc.) full set vs without E2 (ablation)SURFMNet · FAUSTmean geodesic error: 0.077 -> 0.020Improved
add E3 (Laplacian commutativity) to E1+E2+E4 (ablation)SURFMNet · FAUSTmean geodesic error: 0.126 -> 0.020Improved
add E4 to E1+E2+E3 (ablation)SURFMNet · FAUSTmean geodesic error: 0.081 -> 0.020Improved
add E1 to E2+E3+E4 (ablation)SURFMNet · FAUSTmean geodesic error: 0.079 -> 0.020Improved
use E3 alone instead of E4 alone (ablation)SURFMNet · FAUSTmean geodesic error: 0.252 -> 0.073Improved
learn descriptors with the network instead of fixed SHOT descriptors (Fmap Ours Opt vs Fmap Basic) (ablation)Fmap Basic · FAUST remeshed 5kmean geodesic error (x1e-3): 366.2 -> 20.19Improved
add ICP post-processingSURFMNet-sub · FAUST remeshed 5kmean geodesic error (x1e-3): 92.09 -> 30.47Improved
add ICP post-processingSURFMNet-sub · SCAPE remeshed 5kmean geodesic error (x1e-3): 60.32 -> 23.00Improved
add PMF post-processingFMNet · FAUST remeshed 5kmean geodesic error (x1e-3): 112.8 -> 83.61Improved
optimize on all shapes (Ours-all) instead of the 80-shape training subsetSURFMNet · FAUSTmean geodesic error: 0.020 -> 0.008Improved
use unsupervised SURFMNet instead of supervised FMNetSURFMNet · FAUST remeshed 5kmean geodesic error (x1e-3): 112.8 -> 92.09Improved

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