Medical Image Segmentation on LNDb (5-fold cross-validation)
62.4IoUOurs (Full Model w/ SAFT)
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Ours (Full Model w/ SAFT)Tuning Strategy (PEFT/SAFT/Full Tuning)=SAFT, Evaluation Protocol=Fine-Tuning Strategy2026.04 | 62.4 | 74.6 | |
| + Semantic-Topological Graph Reasoning (STGR)Reasoning Module (STGR)=Enabled, Evaluation Protocol=Ablation Strategy2026.04 | 60.8 | 73.2 | |
| Full Model (No PEFT, Full Tuning)Tuning Strategy (PEFT/SAFT/Full Tuning)=Full Tuning, Evaluation Protocol=Fine-Tuning Strategy2026.04 | 60.2 | 72.8 | |
| LISAModel Category=Foundation & Referring Segmentation Models2026.04 | 57.4 | 70.1 | |
| + Text-to-Vision Intent Distillation (TVID)Distillation Strategy (TVID)=Enabled, Evaluation Protocol=Ablation Strategy2026.04 | 57 | 69.4 | |
| BaselinePrompting Strategy=LLaMA-3-V + MedSAM, Evaluation Protocol=Ablation Strategy2026.04 | 55.2 | 67.8 | |
| SegNetModel Category=Standard Medical Segmentation Models2026.04 | 55.1 | 68.4 | |
| SEEMModel Category=Foundation & Referring Segmentation Models2026.04 | 54.9 | 67.5 | |
| TransUNetModel Category=Standard Medical Segmentation Models2026.04 | 53.7 | 66.8 | |
| MedSAMPrompting Strategy=Text-Prompt, Model Category=Foundation & Referring Segmentation Models2026.04 | 51.2 | 65 | |
| U-NetModel Category=Standard Medical Segmentation Models2026.04 | 49.3 | 62.1 |