AI Paper Atlas

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

How to Synthesize a Large-Scale and Trainable Micro-Expression Dataset?

Yuchi Liu, Zhongdao Wang, Tom Gedeon, Liang Zheng

ECCV 2022 · Label-efficient and robust learning › Recognition and long-tail · approach: Uncertainty-aware facial expression recognition · 7 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
pre-train on MiE-X before fine-tuning on CompMiEApexME · CompMiEUF1 and UAR: 5.1% and 7.1% improvementImproved
pre-train on MiE-X instead of MMEWBranches · CompMiEUF1 and UAR: 7.4% and 8.0% higherImproved
pre-train on real MiE datasets (SMIC, CASME, SAMM, MMEW) instead of no pre-trainingBranches · CompMiEUF1: 43.6 -> 42.8Hurt
train directly on MiE-X instead of CompMiE+CK+Branches · CompMiEUF1: 45.2 and 47.7 vs 45.0 and 45.2Improved
add MiE-X pre-trainingApexME · MMEWaccuracy: 48.5 ± 0.6 -> 55.9 ± 2.0Improved
add MiE-X pre-trainingApexME · SAMMaccuracy: 41.3 ± 0.6 -> 46.4 ± 0.7Improved
add MiE-X pre-trainingDTSCNN · MMEWaccuracy: 60.9 ± 1.3 -> 63.1 ± 1.0Improved
add the real-MiE AU type z_MiE to synthesis (ablation)Branches on MiE-X · CompMiEUF1: 1.6% decrease when removedImproved
add the early-stage macro-expression AU type z_MaE to synthesis (ablation)Branches on MiE-X · CompMiEUF1: 1.0% decrease when removedImproved
add the expert-knowledge AU type z_exp to synthesis (ablation)Branches on MiE-X · CompMiEUF1: 1.6% decrease when removedImproved
increase distinct AU triplets from 4,000 to 10,000 (ablation)Branches on MiE-X · CompMiEUF1: +1.8%Improved
increase identities from 700 to 1,700 (ablation)Branches on MiE-X · CompMiEUF1: 44.4% to 45.8%Improved

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