Research map / Label-efficient and robust learning / Domain adaptation and generalization
CXR Segmentation by AdaIN-Based Domain Adaptation and Knowledge Distillation
Yujin Oh, Jong Chul Ye
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 changed | Where | Measured | Effect |
|---|---|---|---|
| add domain adaptation via shared AdaIN generator (Proposed) versus U-Net alone | U-Net · BIMCV-13 abnormal CXR, harsh shift | TPR: 0.82 -> 0.89 | Improved |
| add self-consistency loss (Proposed+l_self) instead of Proposed | Proposed · BIMCV-13 abnormal CXR, harsh shift | TPR: 0.89 -> 0.86 | No clear effect |
| use Proposed instead of XLSor under weak shift | XLSor · BIMCV-13 abnormal CXR, weak shift | TPR: 0.62 -> 0.91 | Improved |
| use Proposed+l_self instead of lungVAE under harsh shift | lungVAE · BIMCV-13 abnormal CXR, harsh shift | TPR: 0.10 -> 0.86 | Improved |
| use Proposed+l_self instead of lungVAE under no shift on normal CXR | lungVAE · NLM normal CXR, no shift | Dice: 0.94 -> 0.91 | Hurt |
| use StarGANv2 style transfer before U-Net | U-Net · BIMCV-13 abnormal CXR, weak shift | TPR: 0.82 -> 0.90 | Improved |
| add segmentation loss to baseline (ablation) | domain adaptation baseline · abnormal CXR (qualitative) | segmentation shape | Mixed |
| add style loss on top of segmentation loss (ablation) | segmentation task · abnormal CXR (qualitative) | segmentation shape | Improved |
| add cycle-consistency loss on top of segmentation and style losses (ablation) | segmentation task · abnormal CXR (qualitative) | segmentation shape | Hurt |
Closest papers by meaning
- Instruction-Guided Lesion Segmentation for Chest X-rays with Automatically Generated Large-Scale Dataset (CVPR 2026 · similarity 0.85)
- ADAS: A Direct Adaptation Strategy for Multi-Target Domain Adaptive Semantic Segmentation (CVPR 2022 · similarity 0.85)
- Unsupervised Intra-Domain Adaptation for Semantic Segmentation Through Self-Supervision (CVPR 2020 · similarity 0.84)
- MAPSeg: Unified Unsupervised Domain Adaptation for Heterogeneous Medical Image Segmentation Based on 3D Masked Autoencoding and Pseudo-Labeling (CVPR 2024 · similarity 0.84)
- SemiDAViL: Semi-supervised Domain Adaptation with Vision-Language Guidance for Semantic Segmentation (CVPR 2025 · similarity 0.83)
- BenchX: A Unified Benchmark Framework for Medical Vision-Language Pretraining on Chest X-Rays (NeurIPS 2024 · similarity 0.83)
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.
