AI Paper Atlas

Research map / Label-efficient and robust learning / Test-time adaptation

Deep Matching Prior: Test-Time Optimization for Dense Correspondence

Sunghwan Hong, Seungryong Kim

ICCV 2021 · Label-efficient and robust learning › Test-time adaptation · approach: Test-time adaptation of vision-language models · 20 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 residual correspondence module (soft-argmax residual) (ablation)DMP baseline · HPatches AEEAEE: 33.1 -> 18.6Improved
add adaptation feature layer alone (ablation)DMP baseline · HPatches AEEAEE: 33.1 -> 32.8No clear effect
add adaptation feature layer on top of residual correspondence (ablation)DMP · HPatches AEEAEE: 18.6 -> 12.3Improved
add confidence threshold to contrastive loss (ablation)DMP · HPatches AEEAEE: 12.3 -> 9.69Improved
increase number of loss samples M from 128 to 256 or 512 (ablation)DMP · HPatches AEEAEE: 9.84 -> 9.69 -> 9.67No clear effect
set confidence threshold too high (ablation)DMP · HPatchesAEEHurt
add test-time augmentation (A-DMP)DMP · HPatches (240x240) AEEAEE: 9.69 -> 8.18Improved
add test-time augmentation (A-DMP)DMP · HPatches original resolution AEEAEE: 30.64 -> 27.22Improved
add RANSAC coarse alignment (RANSAC-DMP)DMP · HPatches (240x240) AEEAEE: 9.69 -> 3.05Improved
add pre-training and test-time optimization (DMP†)DMP · HPatches (240x240) AEEAEE: 9.69 -> 5.21Improved
add test-time optimization to pre-trained DMP†DMP† · HPatches (240x240) AEEAEE: 7.53 -> 5.21Improved
add test-time optimization to pre-trained RANSAC-FlowRANSAC-Flow · HPatches (240x240) AEEAEE: 3.06 -> 5.19Hurt

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