Research map / Detection and segmentation / Open-set and few-shot detection
MRGen: Segmentation Data Engine For Underrepresented MRI Modalities
Haoning Wu, Ziheng Zhao, Ya Zhang, Yanfeng Wang, Weidi Xie
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 MRGen synthetic target-domain data for training | nnUNet · Average over 10 cross-modality pairs | DSC: 10.48 -> 44.71 | Improved |
| add MRGen synthetic data for training | UMamba · Average over 10 cross-modality pairs | DSC: 8.41 -> 43.76 | Improved |
| use MRGen instead of CycleGAN translation | nnUNet · Average over 10 cross-modality pairs | DSC: 26.39 -> 44.71 | Improved |
| use MRGen instead of CycleGAN translation | MRGen vs baselines · Cross-modality generation | FID: 178.18 -> 82.18 | Improved |
| fine-tune SDM on MRGen-DB (ablation) | SDM · MRGen-DB test | FID: 249.24 -> 91.48 | Improved |
| add higher latent dimension and BiomedCLIP text encoder (MRGen-M) (ablation) | SDM-ft · MRGen-DB test | FID: 91.48 -> 41.82 | Improved |
| add templated text prompts with attributes, regions and organs (ablation) | MRGen-M · MRGen-DB test | FID: 41.82 -> 39.63 | Improved |
| add templated text prompts (ablation) | MRGen-M · MRGen-DB test | CLIP-I: 0.7512 -> 0.8457 | Improved |
| add AutoFilter sample selection (ablation) | MRGen + nnUNet · CHAOS-MRI T1->T2-SPIR | DSC: 16.53 -> 22.30 | Improved |
| add unannotated target-domain images (ablation) | MRGen + nnUNet · CHAOS-MRI T1->T2-SPIR | DSC: 16.53 -> 30.16 | Improved |
| add both AutoFilter and target-domain images (ablation) | MRGen + nnUNet · CHAOS-MRI T1->T2-SPIR | DSC: 16.53 -> 66.18 | Improved |
| add both AutoFilter and target-domain images (ablation) | MRGen + nnUNet · MSD-Prostate ADC->T2 | DSC: 18.92 -> 61.95 | Improved |
Closest papers by meaning
- Generative Medical Segmentation (AAAI 2025 · similarity 0.89)
- Training Like a Medical Resident: Context-Prior Learning Toward Universal Medical Image Segmentation (CVPR 2024 · similarity 0.89)
- MAPSeg: Unified Unsupervised Domain Adaptation for Heterogeneous Medical Image Segmentation Based on 3D Masked Autoencoding and Pseudo-Labeling (CVPR 2024 · similarity 0.87)
- LangDAug: Langevin Data Augmentation for Multi-Source Domain Generalization in Medical Image Segmentation (ICML 2025 · similarity 0.87)
- Enhancing SAM with Efficient Prompting and Preference Optimization for Semi-supervised Medical Image Segmentation (CVPR 2025 · similarity 0.87)
- Interactive Medical Image Segmentation: A Benchmark Dataset and Baseline (CVPR 2025 · similarity 0.86)
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.
