Egocentric Referring Video Object Segmentation on VISOR (val)
67mIoUCERES
Evaluation results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| CERESBackbone=SwinL2025.12 | 67 | 73.6 | 16.9 | 28.6 | 71.8 | 75.2 | 76.2 | |
| ActionVOSBackbone=SwinL2025.12 | 66.3 | 71.9 | 22.8 | 42.5 | 68.7 | 73.4 | 75.5 | |
| CERESBackbone=VSwinB2025.12 | 65.4 | 72.5 | 19.1 | 35.1 | 72.1 | 74.7 | 75.9 | |
| CERESBackbone=R1012025.12 | 64 | 72.8 | 15.3 | 25.6 | 72.4 | 76.3 | 77.1 | |
| ActionVOSBackbone=VSwinB2025.12 | 62.9 | 70.9 | 20 | 38.8 | 69.5 | 70.7 | 74.3 | |
| ReferFormer+Backbone=VSwinB2025.12 | 61.1 | 68.9 | 19.2 | 36.8 | 68.4 | 73.2 | 74 | |
| ReferFormerBackbone=R1012025.12 | 59.9 | 66.4 | 30.5 | 52.1 | 55.3 | 58.6 | 64.2 | |
| ActionVOSBackbone=R1012025.12 | 59.9 | 67.2 | 16.3 | 28.5 | 69.9 | 73.4 | 73.7 | |
| ReferFormer+Backbone=R1012025.12 | 58.2 | 64.8 | 14.3 | 18.9 | 63.1 | 67.6 | 68.9 | |
| HOSBackbone=SwinL2025.12 | 55.1 | 59.2 | 13.5 | 17.3 | 66.5 | 70.3 | 69.7 |