AI Paper Atlas

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

Rethinking Semi-Supervised Medical Image Segmentation: A Variance-Reduction Perspective

Chenyu You, Weicheng Dai, Yifei Min, Fenglin Liu, David A. Clifton, S Kevin Zhou, Lawrence Hamilton Staib, James s Duncan

NeurIPS 2023 · Label-efficient and robust learning › Semi-supervised learning · approach: Semi-supervised medical segmentation · 62 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 global contrastive (tailness) loss (ablation)ARCO-SG · ACDC 1% labelsaverage Dice (%): 60.9 -> 85.5Improved
add nearest-neighbor diversity loss (ablation)ARCO-SG · ACDC 1% labelsaverage Dice (%): 79.3 -> 85.5Improved
add nearest-neighbor diversity loss (ablation)ARCO-SAG · ACDC 1% labelsaverage Dice (%): 78.6 -> 84.9Improved
use stratified group sampling instead of naive sampling (ablation)ARCO · ACDC 1% labelsaverage Dice (%): 82.6 -> 85.5Improved
use stratified-antithetic group sampling instead of naive sampling (ablation)ARCO · ACDC 1% labelsaverage Dice (%): 82.6 -> 84.9Improved
add unsupervised loss (ablation)ARCO-SG · ACDC 1% labelsaverage Dice (%): 81.9 -> 85.5Improved
add global instance discrimination loss (ablation)ARCO-SG · ACDC 1% labelsaverage Dice (%): 84.1 -> 85.5Improved
add local instance discrimination loss (ablation)ARCO-SG · ACDC 1% labelsaverage Dice (%): 83.8 -> 85.5Improved
use ARCO-SG instead of MONAcontrastive semi-supervised segmentation · ACDC 1% labelsaverage Dice (%): 82.6 -> 85.5Improved
use ARCO-SAG instead of ARCO-SGARCO · ACDC 1% labelsaverage Dice (%): 84.9 -> 85.5 Dice at 1%Mixed
add data augmentation strategies (ablation)ARCO · ACDC 1% labelsaverage Dice (%)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