Segmentation on MIMOSeg Perspective IV held-in (test)
58.6mIoUMIMO
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| MIMOMulti-modal Input Aligner=true, LLM=Vicuna-7B2025.10 | 58.6 | 30.9 | 0.47 | |
| SAM-hZero-shot=true, Backbone=ViT-H2025.10 | 58.3 | 40.4 | — | |
| SAM-bZero-shot=true, Backbone=ViT-B2025.10 | 57.3 | 36.3 | — | |
| GLaMMFine-tuned=true2025.10 | 56.4 | 26.7 | 0.422 | |
| MIMO(w/o Aligner)Multi-modal Input Aligner=false, LLM=Vicuna-7B2025.10 | 52.6 | 25.3 | 0.401 | |
| SAM-MedZero-shot=true2025.10 | 50.3 | 30 | — |