AI Paper Atlas

Research map / 3D vision / Face and shape

Do It Yourself: Learning Semantic Correspondence from Pseudo-Labels

Olaf Dünkel, Thomas Wimmer, Christian Theobalt, Christian Rupprecht, Adam Kortylewski

ICCV 2025 · 3D vision › Face and shape · approach: Optimization-based shape matching · 10 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
add pseudo-label adapter training (naive labels) over zero-shot (ablation)SD+DINOv2 · SPair-71kPCK@0.1: 65.0 -> 67.2Improved
add exact cyclic consistency filtering (ablation)pseudo-labels · SPair-71kPCK@0.1: 67.2 -> 66.9Hurt
add relaxed cyclic consistency filtering (ablation)pseudo-labels · SPair-71kPCK@0.1: 67.2 -> 68.4Improved
add 3D-aware chaining (ablation)pseudo-labels with relaxed cyclic consistency · SPair-71kPCK@0.1: 68.4 -> 70.0Improved
add spherical prototype rejection (ablation)pseudo-labels with relaxed cyclic consistency and chaining · SPair-71kPCK@0.1: 70.0 -> 74.4Improved
add spherical rejection to naive pseudo-labels (ablation)naive pseudo-labels · SPair-71kPCK@0.1: 67.2 -> 72.9Improved
use ground-truth pose instead of Orient-Anything estimated pose (ablation)DIY-SC · SPair-71k (per-image)PCK@0.1: 69.6 -> 71.6Improved
use SAM masks instead of rembg masks (ablation)DIY-SC with estimated pose · SPair-71k (per-image)PCK@0.1: 68.0 -> 69.6Improved
generate pseudo-labels with SD+DINOv2 instead of DINOv2 only (ablation)DINOv2 refiner · SPair-71k (per-image)PCK@0.1: 67.17 -> 71.6Improved
add ImageNet-3D pretraining then fine-tune on SPairDIY-SC · SPair-71k (per-image)PCK@0.1: 71.6 -> 72.2Improved
reduce adapter output channels from 1536 to 128 (ablation)DIY-SC · SPair-71kPCK@0.1: 74.63 -> 74.08No clear effect
fine-tune adapter with ground-truth labels after pseudo-label pretrainingsupervised TLR · SPair-71k (per-image, supervised)PCK@0.1: 82.9 -> 83.5Improved

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