AI Paper Atlas

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

Improving Semantic Segmentation via Efficient Self-Training

Yi Zhu, Zhongyue Zhang, Chongruo Wu, Zhi Zhang, Tong He, Hang Zhang, Raghaven Manmatha, Mu Li, Alexander Johannes Smola

TPAMI 2021 · Label-efficient and robust learning › Semi-supervised learning · approach: Consistency training for semi-supervised segmentation · 154 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 pseudo labels from Cityscapes coarse imagesSelf-training student · Cityscapes valmIoU: 78.1 -> 79.0Improved
add pseudo labels from Mapillary images as wellSelf-training student · Cityscapes valmIoU: 78.1 -> 79.3Improved
add centroid sampling (ablation)Self-training student · Cityscapes valmIoU: 78.9 -> 80.0Improved
increase pseudo-to-real ratio from 1:1 to 7:1 (ablation)Self-training student with centroid sampling · Cityscapes valmIoU: 79.3 -> 80.0Improved
increase pseudo ratio to 7:1 without centroid sampling (ablation)Self-training student without centroid sampling · Cityscapes valmIoU: 78.9 -> 78.7Hurt
replace real-label duplication with pseudo labels (ablation)Self-training student · Cityscapes valmIoU: 79.1 -> 80.0Improved
use self-training instead of Mapillary labeled pre-trainingDeepLabV3+ WideResNet38 · Cityscapes valmIoU: 81.5 -> 82.2Improved
add warm-up with crop 800 for first 20 epochs to coarse2fine+ schedule (ablation)Fast training schedule · Cityscapes valmIoU: 79.5 -> 80.0Improved
use soft pseudo labels instead of hard (ablation)Self-training student · Cityscapes valmIoU: 80.0 -> 79.6Hurt
pre-train on pseudo labels then finetune instead of joint learning (ablation)Self-training student · Cityscapes valmIoU: 80.0 -> 79.2Hurt
add extra teacher-student loops (2-loop, 3-loop) (ablation)Self-training student · Cityscapes valmIoU: 80.0 -> 79.9 -> 80.2No clear effect
use a better teacher (WideResNet38, 80.5 vs 78.1) (ablation)Student models · Cityscapes valmIoU: +0.2 to +0.5Improved

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