AI Paper Atlas

Research map / Label-efficient and robust learning / Self-supervised and contrastive

Beyond Instance-Level Self-Supervision in 3D Multi-Modal Medical Imaging

Tan Pan, Shuhao Mei, Yixuan Sun, Kaiyu Guo, Chen Jiang, Zhaorui Tan, Mengzhu Li, LIMEI HAN, Xiang Zou, Yuan Cheng, Mahsa Baktashmotlagh

ICML 2026 · Label-efficient and robust learning › Self-supervised and contrastive · approach: Self-supervised neuroimaging representations · 0 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 Lintra and Linter topology-consistency losses to reconstruction pre-training (ablation)Luni baseline · BraTS-METavg Dice (%): 61.63 -> 68.17Improved
add Lintra and Linter losses to reconstruction pre-training (ablation)Luni baseline · ADHD-200F1 (%): 63.06 -> 63.82Improved
add Lintra and Linter losses to reconstruction pre-training (ablation)Luni baseline · ADHD-200ACC (%): 64.94 -> 67.53Improved
add only one of Lintra or Linter (ablation)Luni baseline · ADHD-200F1 (%): 63.06 -> 57.61 (first row), 63.06 -> 61.44 (second row)Mixed
switch from Swin UNETR-based BrainMVP to TACO pre-trainingsegmentation pre-training · four segmentation datasets averageDice (%): 76.36 -> 77.47Improved
pre-train with TACO instead of best baselines on classificationSwin-B on ADNI and ADHD · average F1F1 (%): 72.76 -> 78.7Improved
TACO pre-training on ADHD-200 accuracySwin-B · ADHD-200ACC (%): 69.26 -> 67.53Hurt
use TACO pre-training for missing-modality segmentationBraTS18 segmentation · whole tumor, all modality configurationsmean Dice (%): 87.63 -> 89.55Improved
use TACO pre-training for missing-modality segmentationBraTS18 segmentation · tumor core, all modality configurationsmean Dice (%): 79.78 -> 79.68No clear effect
increase neighbourhood size Top-omega (ablation)TACO · BraTS-MET and ADHDdownstream performanceMixed
increase triplet margin delta (ablation)TACO · BraTS-MET and ADHDdownstream performanceMixed
TACO pre-training with fewer labeled samples (5-10%)BraTS-MET segmentation and ADHD classification · label efficiencyperformanceImproved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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