AI Paper Atlas

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

Deep Imbalanced Learning for Face Recognition and Attribute Prediction

Chen Huang, Yining Li, Chen Change Loy, Xiaoou Tang

TPAMI 2019 · Label-efficient and robust learning › Recognition and long-tail · approach: Classifier rebalancing and margin losses · 362 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 uniform-to-hardest query cluster sampling (choose highest-loss cluster instead of uniform) (ablation)CLMLE · LFWaccuracy (%): 99.52 -> 99.62Improved
add highest-loss query cluster sampling instead of uniform (ablation)CLMLE · CelebAaverage balanced accuracy (%): 88.23 -> 88.78Improved
add cost-sensitive learning (inverse class frequency in batch) (ablation)CLMLE · LFWaccuracy (%): 99.57 -> 99.62Improved
add cost-sensitive learning (inverse class frequency in batch) (ablation)CLMLE · CelebAaverage balanced accuracy (%): 87.46 -> 88.78Improved
use k-nearest cluster classifier instead of instance-wise kNN (ablation)CLMLE · CelebAaverage balanced accuracy (%): 88.59 -> 88.78Improved
replace prior features for initial clustering (softmax vs triplet pretraining) (ablation)CLMLE · LFWaccuracy (%): no more than 0.02%No clear effect
replace prior features for initial clustering (face recognition vs multi-attribute softmax) (ablation)CLMLE · CelebAaccuracy (%): no more than 0.3%No clear effect
replace quintuplet sampling (LMLE) with entire cluster sampling (CLMLE)LMLE · CelebAaverage balanced accuracy (%): 83.83 -> 88.78Improved
replace quintuplet sampling (LMLE) with cluster sampling (CLMLE)LMLE · MegaFace Challenge 1 (small protocol)verification TAR (%): 89.45 -> 91.85Improved
use CLMLE instead of CE+CRL on imbalanced CIFAR-100CE+CRL · CIFAR-100 (gamma=0.5)mean class-balanced accuracy (%): 37.4 -> 39.7Improved
use CLMLE instead of CE+CRL on imbalanced MNIST-rot-back-imageCE+CRL · MNIST-rot-back-image (gamma=0.5)mean class-balanced accuracy (%): 64.5 -> 68.8Improved
use CLMLE instead of Triplet+ on imbalanced CIFAR-100Triplet+ · CIFAR-100 (gamma=0.5)mean class-balanced accuracy (%): 32.2 -> 39.7Improved

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