Open-vocabulary Segmentation on PC59 OV
64.6mIoUHyperSeg
Evaluation results
| Method | Links | |
|---|---|---|
| HyperSegType=MLLM based Segmentation Network, Backbone=Swin-B2024.11 | 64.6 | |
| CAT-SegPrompt Tuning=MMRL+VaMP2025.11 | 58.6 | |
| CAT-SegPrompt Tuning=MMRL2025.11 | 57.9 | |
| CAT-SegPrompt Tuning=-2025.11 | 57.5 | |
| SANPrompt Tuning=-2025.11 | 53.8 | |
| OVSegPrompt Tuning=-2025.11 | 53.3 | |
| DaTaSegType=Segmentation Specialist, Backbone=ViTDet-B2024.11 | 51.1 | |
| DeOPType=Segmentation Specialist, Backbone=ResNet-101c2024.11 | 48.8 | |
| PSALMType=MLLM based Segmentation Network, Backbone=Swin-B2024.11 | 48.5 | |
| SimBaselineType=Segmentation Specialist, Backbone=ViT-B2024.11 | 47.7 | |
| MaskCLIPType=Segmentation Specialist, Backbone=ViT-L2024.11 | 45.9 | |
| ZegFormerPrompt Tuning=-2025.11 | 45.5 |