AI Paper Atlas

Research map / Label-efficient and robust learning / Domain adaptation and generalization

CoSMIC: Continual Self-supervised Learning for Multi-Domain Medical Imaging via Conditional Mutual Information Maximization

Yihang Liu, Ying Wen, Longzhen Yang, Lianghua He, Heng Tao Shen

ICCV 2025 · Label-efficient and robust learning › Domain adaptation and generalization · approach: Augmentation and mixture-of-experts generalization · 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 reconstruction loss L_Recon (ablation)CoSMIC · Retinoscopy D1A (avg): 85.79 -> 87.70Improved
add the ([XZ],Y) similarity term in L_CoCst (ablation)CoSMIC · Retinoscopy D1A (avg): 78.64 -> 87.70Improved
add the (X,Y) similarity term in L_CoCst (ablation)CoSMIC · Retinoscopy D1A (avg): 83.63 -> 87.70Improved
add Anatomy-Guided Calibration (ablation)CoSMIC · Retinoscopy D1A (avg): 85.04 -> 87.70Improved
add Anatomy-Guided Calibration (ablation)CoSMIC · Retinoscopy D1KB: 0.91 -> 0.21Improved
add Hierarchical Alignment (ablation)CoSMIC · Retinoscopy D1A (avg): 86.59 -> 87.70Improved
add the ([XZ],Y) similarity term in L_CoCst (ablation)CoSMIC · Retinoscopy D1KB: 8.07 -> 0.21Improved
add the ([XZ],Y) similarity term in L_CoCst (ablation)CoSMIC · Histopathology D5 (unseen)KF: -9.43 -> -0.71Improved
reverse pre-training order to D4-D3-D2-D1 (ablation)CoSMIC · Retinoscopy D1A (avg): 87.70 -> 87.32No clear effect
use CoSMIC instead of MedCoss (rehearsal) when moving from D0 to D1continual SSL · D1 knowledge backwardKB: 7.33 -> 0.21Improved
use CoSMIC instead of RETFound on APTOSfundus pre-training · APTOSAUC (%): 94.48 -> 95.39Improved
use CoSMIC instead of RETFound on PAPILAfundus pre-training · PAPILAAUC (%): 84.51 -> 80.10Hurt

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