Research map / Label-efficient and robust learning / Domain adaptation and generalization
Reducing Domain Gap in Frequency and Spatial Domain for Cross-Modality Domain Adaptation on Medical Image Segmentation
Shaolei Liu, Siqi Yin, Linhao Qu, Manning Wang
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 |
|---|---|---|---|
| combine frequency and spatial transfer teachers through multi-teacher distillation (ablation) | TransUNet UDA · Abdominal multi-organ | Dice: 88.4 -> 89.2 | Improved |
| combine frequency and spatial transfer teachers through multi-teacher distillation (ablation) | TransUNet UDA · Cardiac | Dice: 83.9 -> 85.3 | Improved |
| use frequency transfer instead of no adaptation (ablation) | TransUNet UDA · Abdominal multi-organ | Dice: 53.3 -> 87.3 | Improved |
| use frequency transfer instead of no adaptation (ablation) | TransUNet UDA · Cardiac | Dice: 46.0 -> 84.6 | Improved |
| use spatial histogram transfer instead of no adaptation (ablation) | TransUNet UDA · Abdominal multi-organ | Dice: 53.3 -> 88.6 | Improved |
| use spatial histogram transfer instead of no adaptation (ablation) | TransUNet UDA · Cardiac | Dice: 46.0 -> 83.7 | Improved |
| mix spatial- and frequency-transferred images in one model instead of using multi-teacher distillation (ablation) | TransUNet UDA · Abdominal multi-organ | Dice: 89.2 -> 88.4 | Hurt |
| mix spatial- and frequency-transferred images in one model instead of using multi-teacher distillation (ablation) | TransUNet UDA · Cardiac | Dice: 85.3 -> 83.9 | Hurt |
| use batch momentum histogram matching instead of whole-domain histogram matching (ablation) | Spatial transfer · Cardiac | Dice: 81.5 -> 83.7 | Improved |
| use batch momentum histogram matching instead of single-image histogram matching (ablation) | Spatial transfer · Cardiac | Dice: 82.8 -> 83.7 | Improved |
| use batch momentum histogram matching instead of current-batch average matching (ablation) | Spatial transfer · Cardiac | Dice: 83.4 -> 83.7 | Improved |
| retain low-frequency FC [0,1] and third-layer FC [7,14], replacing other source frequency components with target components (ablation) | NSCT frequency transfer · Cardiac synthetic target validation | Dice: 60.42 -> 64.85 | Improved |
Closest papers by meaning
- One-to-Multiple: A Progressive Style Transfer Unsupervised Domain-Adaptive Framework for Kidney Tumor Segmentation (NeurIPS 2024 · similarity 0.88)
- A Unified Degradation-Robust Approach to SSL and UDA for 3D Medical Images (AAAI 2025 · similarity 0.88)
- MAPSeg: Unified Unsupervised Domain Adaptation for Heterogeneous Medical Image Segmentation Based on 3D Masked Autoencoding and Pseudo-Labeling (CVPR 2024 · similarity 0.88)
- Universal Domain Adaptation for Semantic Segmentation (CVPR 2025 · similarity 0.87)
- AIF-SFDA: Autonomous Information Filter Driven Source-Free Domain Adaptation for Medical Image Segmentation (AAAI 2025 · similarity 0.87)
- FAMNet: Frequency-aware Matching Network for Cross-domain Few-shot Medical Image Segmentation (AAAI 2025 · similarity 0.87)
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.
