Referring Video Object Segmentation on JHMDB (val)
43.7mIoUReferFormer
Evaluation results
| Method | Links | |
|---|---|---|
| ReferFormerVis. Backbone=Vid-Swin-B, Eval=Fine-Tuned, Label-Rate=100%2024.07 | 43.7 | |
| SAFARIVis. Backbone=Swin-B, Eval=Zero-Shot, Label-Rate=100%2024.07 | 43.2 | |
| PolyFormerVis. Backbone=Swin-B, Eval=Zero-Shot, Label-Rate=100%2024.07 | 42.4 | |
| SAFARIVis. Backbone=Swin-B, Eval=Zero-Shot, Label-Rate=30%2024.07 | 38.1 | |
| MTTRVis. Backbone=Vid-Swin-T, Eval=Fine-Tuned, Label-Rate=100%2024.07 | 36.6 | |
| SAFARIVis. Backbone=Swin-B, Eval=Zero-Shot, Label-Rate=10%2024.07 | 36.4 | |
| SeqTRVis. Backbone=DN-53, Eval=Zero-Shot, Label-Rate=100%2024.07 | 34.9 | |
| Partial-RESVis. Backbone=Swin-B, Eval=Zero-Shot, Label-Rate=30%2024.07 | 34.8 | |
| Partial-RESVis. Backbone=Swin-B, Eval=Zero-Shot, Label-Rate=10%2024.07 | 34.3 |