AI Paper Atlas

Research map / 3D vision / Face and shape

Unsupervised Deep Multi-Shape Matching

Dongliang Cao, Florian Bernard

ECCV 2022 · 3D vision › Face and shape · approach: Optimization-based shape matching · 41 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
remove the universe classifier and convert functional maps to point maps by nearest neighbour (classifier-free) (ablation)Unsupervised multi-shape matching with universe classifier · FAUST/SCAPE incl. cross-dataset generalisationmean geodesic errorImproved
add the universe classifier instead of soft point maps from feature similarity with Sinkhorn (ablation)Unsupervised multi-shape matching with universe classifier · FAUST/SCAPEmean geodesic errorImproved
use feature-similarity point maps rather than nearest-neighbour conversion of functional maps (ablation)classifier-free ablation · cross-dataset generalisationmean geodesic errorImproved
replace the unsupervised functional-map loss with a supervised cross-entropy loss on ground truth correspondences (ablation)Unsupervised multi-shape matching with universe classifier · FAUST vs other datasets and generalisationmean geodesic errorMixed
fine-tune the pretrained network per test shape pair with the unsupervised lossUnsupervised multi-shape matching · cross-dataset generalisation (FAUST/SCAPE)mean geodesic errorImproved

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