Referring Image Segmentation on OmniRef (test)
69.27cIoUOmniSegNet
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| OmniSegNetBackbone=Swin-B, Text Encoder=BERT2025.12 | 69.27 | 67.8 | 57.69 | |
| DCAMA+ReLABackbone=Swin-B, Text Encoder=BERT2025.12 | 62.61 | 58.08 | 55.93 | |
| VRP-SAM+ReLABackbone=ViT-L, Text Encoder=BERT2025.12 | 61.32 | 59.07 | 55.71 |