Research map / Detection and segmentation / Open-set and few-shot detection
Medverse: A Universal Model for Full-Resolution 3D Medical Image Segmentation, Transformation and Enhancement
Jiesi Hu, Jianfeng Cao, Yanwu Yang, Chenfei Ye, Yixuan Zhang, Hanyang Peng, Ting Ma
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 |
|---|---|---|---|
| use Medverse instead of Neuroverse3D | Neuroverse3D · Held-out 3D segmentation, average (context size 4) | Dice (%): 81.10 -> 87.27 | Improved |
| use Medverse instead of Neuroverse3D on unseen kidney | Neuroverse3D · Held-out segmentation: Kidney (unseen organ) | Dice (%): 75.75 -> 95.31 | Improved |
| use Medverse instead of Neuroverse3D | Neuroverse3D · Held-out transformation and enhancement, average (context size 4) | PSNR (dB): 27.08 -> 28.30 | Improved |
| use Medverse instead of few-shot 3D U-Net | 3D U-Net · Held-out transformation and enhancement, average (context size 4) | PSNR (dB): 29.16 -> 28.30 | Hurt |
| use Medverse instead of few-shot Swin-UNETR (no fine-tuning) | Swin-UNETR · Held-out 3D segmentation, average (context size 4) | Dice (%): 87.67 -> 87.27 | No clear effect |
| add NA-ICL next-scale autoregressive framework (vs sliding window) (ablation) | Medverse · Ablation (BAM and NA-ICL) | Seg. Dice (%): 76.50/78.90 -> 87.27 | Improved |
| add BAM blockwise cross-attention (vs plain concatenation) (ablation) | Medverse · Ablation (BAM and NA-ICL) | Seg. Dice (%): 76.50/78.90 -> 87.27 | Improved |
| increase context size from 1 to 16 (ablation) | Medverse · Held-out segmentation, context size 1 to 16 | Dice gain (points): +7.28 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- SegGPT — SegGPT: Towards Segmenting Everything in Context (ICCV 2023) · better on 1 of 1 shared result
- UniverSeg — UniverSeg: Universal Medical Image Segmentation (ICCV 2023) · better on 1 of 1 shared result
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
- Training Like a Medical Resident: Context-Prior Learning Toward Universal Medical Image Segmentation (CVPR 2024 · similarity 0.89)
- UniPro: Unified Multi-Mode Medical Image Segmentation from 2D Images to 3D Volumes via Propagation (NeurIPS 2026 · similarity 0.88)
- Autoregressive Sequence Modeling for 3D Medical Image Representation (AAAI 2025 · similarity 0.88)
- VISTA3D: A Unified Segmentation Foundation Model For 3D Medical Imaging (CVPR 2025 · similarity 0.87)
- Better Tokens for Better 3D: Advancing Vision-Language Modeling in 3D Medical Imaging (NeurIPS 2025 · similarity 0.87)
- Show and Segment: Universal Medical Image Segmentation via In-Context Learning (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.
