Spatio-temporal video grounding on VidSTG (test) (Declarative & Interrogative)
34.8Declarative m_vIoUST-GD
Evaluation Results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| ST-GDBackbone=Swin-B, Training Method=PEFT2026.04 | 34.8 | 51.5 | 48.5 | 33.4 | 51.2 | 30.3 | 41.2 | 28.8 | |
| VideoGrounding-DINOTraining Method=Partially Fine-tuned2026.04 | 34.6 | 51.9 | 48.1 | 33.9 | 50.8 | 29.8 | 41 | 27.6 | |
| CG-STVGTraining Method=Trained from Scratch2026.04 | 34 | 51.4 | 47.7 | 33.1 | 49.9 | 29 | 40.5 | 27.5 | |
| CSDVLTraining Method=Trained from Scratch2026.04 | 33.7 | — | 47.2 | 32.8 | — | 28.5 | 39.9 | 26.2 | |
| ST-GDBackbone=Swin-T, Training Method=PEFT2026.04 | 33.6 | 50.2 | 45.7 | 32.8 | 49.6 | 28.7 | 39.4 | 27.3 | |
| STCATTraining Method=Trained from Scratch2026.04 | 33.1 | 50.8 | 46.2 | 32.6 | 49.7 | 28.2 | 39.2 | 26.6 | |
| TubeDETRTraining Method=Trained from Scratch2026.04 | 30.4 | 48.1 | 42.5 | 28.2 | 46.9 | 25.7 | 35.7 | 23.2 | |
| RealVGTraining Method=Zero Shot2026.04 | 29 | 37.4 | 34.9 | 25.5 | 36.7 | 30 | 34.7 | 25.7 | |
| CVTPTraining Method=Weakly Supervised2026.04 | 17.9 | — | 22.3 | 14.9 | — | 11.2 | 12.4 | 7.2 | |
| E3MPTraining Method=Zero Shot2026.04 | 16.2 | — | 20.4 | 11.9 | — | 10.6 | 12.2 | 5.4 | |
| STProTraining Method=Weakly Supervised2026.04 | 15.5 | — | 19.4 | 12.6 | — | 12.5 | 14.9 | 9.3 |