AI Paper Atlas

Research map / Label-efficient and robust learning / Recognition and long-tail

The Met Dataset: Instance-level Recognition for Artworks

Nikolaos-Antonios Ypsilantis, Noa Garcia, Guangxing Han, Sarah Ibrahimi, Nanne Van Noord, Giorgos Tolias

NeurIPS 2021 · Label-efficient and robust learning › Recognition and long-tail · approach: Loss design and classifier theory · 15 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 PCA whitening and multi-scale (ablation)R18IN kNN · Met dataset testGAP: 3.7 -> 15.9Improved
use softmax confidence normalization (k=3 instead of k=1) (ablation)R18IN kNN · Met dataset testGAP: 2.9 -> 15.9Improved
include distractors in validation for tuning k and tau (ablation)R18IN kNN · Met dataset testGAP: 14.1 -> 15.9Improved
kNN on backbone instead of parametric classifier (AF loss)R18IN DNet AF · Met dataset testGAP: 16.9 -> 23.7Improved
ArcFace loss instead of cross-entropyR18IN DNet · Met dataset testGAP: 21.6 -> 23.7Improved
self-supervised SimSiam training on MetR18IN kNN · Met dataset testGAP: 15.9 -> 26.8Improved
add negatives to contrastive training (Con-Syn vs SimSiam)R18IN · Met dataset testGAP: 26.8 -> 30.4Improved
choose closest real positive instead of random real positive (ablation)R18IN Con-Syn+Real · Met dataset testGAP: 29.8 -> 32.5Improved
add random real positives to synthetic positives (ablation)R18IN Con-Syn · Met dataset testGAP: 30.4 -> 29.8Hurt
train on Met with contrastive loss starting from SWSLR18SWSL · Met dataset testGAP: 24.7 -> 36.1Improved
pretrain on SemArt art-attribute prediction instead of ImageNetR50 kNN · Met dataset testGAP: 22.2 -> 1.8Hurt
SWSL pretraining instead of ImageNetR50 kNN · Met dataset testGAP: 22.2 -> 30.4Improved

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