Research map / Detection and segmentation / Open-set and few-shot detection
R2-Seg: Training-Free OOD Medical Tumor Segmentation via Anatomical Reasoning and Statistical Rejection
Shuaike Shen, Ke Liu, Jiaqing Xie, Shangde Gao, Chunhua Shen, Ge Liu, Mireia Crispin-Ortuzar, Shangqi Gao
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 statistical two-sample test and FP gating on top of the ablated pipeline (ablation) | R2-Seg · Bladder tumor (OOD, MR) | Dice: 0.099 -> 0.297 | Improved |
| add statistical two-sample test and FP gating on top of the ablated pipeline (ablation) | R2-Seg · Prostate tumor (OOD, MR) | Dice: 0.069 -> 0.465 | Improved |
| add statistical two-sample test and FP gating on top of the ablated pipeline (ablation) | R2-Seg · Bladder tumor (OOD, MR) | specificity: 0.089 -> 0.536 | Improved |
| enable full pipeline versus only one component (ablation) | R2-Seg · Bladder tumor (OOD, MR) | sensitivity: 0.923 -> 0.335 | Hurt |
| use R2-Seg instead of zero-shot BiomedParse | BiomedParse · Bladder tumor (OOD, MR) | specificity: 0.000 -> 0.536 | Improved |
| use R2-Seg instead of zero-shot BiomedParse | BiomedParse · Prostate tumor (OOD, MR) | Dice: 0.047 -> 0.465 | Improved |
| fine-tune with LoRA instead of training-free R2-Seg | BiomedParse · Bladder tumor (OOD, MR) | Dice: 0.297 -> 0.578 | Improved |
| fine-tuning BiomedParse on tumor slices | BiomedParse · in-distribution organ segmentation (AMOS22 CT/MR) | organ segmentation performance | Hurt |
Closest papers by meaning
- SegVol: Universal and Interactive Volumetric Medical Image Segmentation (NeurIPS 2024 · similarity 0.85)
- Advancing Generalizable Tumor Segmentation with Anomaly-Aware Open-Vocabulary Attention Maps and Frozen Foundation Diffusion Models (CVPR 2025 · similarity 0.85)
- CG-Reasoner: Centroid-Guided Positional Reasoning Segmentation for Medical Imaging with a Robust Visual-Text Consistency Metric (CVPR 2026 · similarity 0.85)
- ZePT: Zero-Shot Pan-Tumor Segmentation via Query-Disentangling and Self-Prompting (CVPR 2024 · similarity 0.85)
- "Why Is There a Tumor?": Tell Me the Reason, Show Me the Evidence (ICML 2025 · similarity 0.84)
- S2S2: Semantic Stacking for Robust Semantic Segmentation in Medical Imaging (AAAI 2025 · similarity 0.84)
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.
