AI Paper Atlas

Research map / Label-efficient and robust learning / Semi-supervised learning

Transductive Centroid Projection for Semi-supervised Large-scale Recognition

Yu Liu, Guanglu Song, Jing Shao, Xiao Jin, Xiaogang Wang

ECCV 2018 · Label-efficient and robust learning › Semi-supervised learning · approach: Holistic semi-supervised pipelines · 34 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 self-training on clustered unlabeled data (ablation)SU (cluster IDs only) · Market-1501Top-1 accuracy: 22.8 -> 65.0Improved
add self-training to the labeled+unlabeled combination (ablation)MU+L · Market-1501Top-1 accuracy: 37.4 -> 68.8Improved
replace naive combination with TCP on unlabeled data (ablation)MU+L · Market-1501Top-1 accuracy: 37.4 -> 89.6Improved
use real annotated ground truth instead of pseudo cluster IDs for unlabeled data (ablation)SU · Market-1501Top-1 accuracy: 22.8 -> 66.4Improved
add manually annotated unlabeled data to labeled training (ablation)SL · Market-1501Top-1 accuracy: 87.7 -> 86.0Hurt
add naive unlabeled cluster data to labeled data (ablation)SL · Market-1501Top-1 accuracy: 87.7 -> 37.4Hurt
add triplet loss on unlabeled data (ablation)MU+L · Market-1501Top-1 accuracy: 37.4 -> 83.5Improved
use TCP instead of triplet loss on unlabeled data (ablation)MU+L triplet loss · Market-1501Top-1 accuracy: 83.5 -> 89.6Improved
use TCP instead of self-training (ablation)MU+L self-training · Market-1501Top-1 accuracy: 68.8 -> 89.6Improved
add TCP semi-supervised training to the labeled baseline (ablation)SL · Market-1501Top-1 accuracy: 87.7 -> 90.4Improved
add re-ranking to TCPTCP · Market-1501mAP: 66.2 -> 84.4Improved
add TCP unlabeled-data training to baselineBaseline · CUHK03Top-1 accuracy: 91.7 -> 94.4Improved

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