Spatio-temporal video grounding on HC-STVG v2
60.4m_tIoUTA-STVG
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| TA-STVGSetting=Fully Sup2025.12 | 60.4 | 40.2 | 65.8 | 36.7 | |
| CG-STVGSetting=Fully Sup2025.12 | 60 | 39.5 | 64.5 | 36.3 | |
| STVGFormerSetting=Fully Sup2025.12 | 58.1 | 38.7 | 65.5 | 33.8 | |
| DEViLZero-shot=false, Time=3.50h2025.12 | 58 | 36.5 | 58 | 31.8 | |
| DEViLSetting=LLM2025.12 | 58 | 36.5 | 58 | 31.8 | |
| TubeDETRSetting=Fully Sup2025.12 | 53.9 | 36.4 | 58.8 | 30.6 | |
| DEViLZero-shot=true, Time=3.50h2025.12 | 52.9 | 28 | 43.8 | 21.1 | |
| LLaVA-STZero-shot=true, Time=4.02h2025.12 | 21.6 | 8.7 | 6 | 1.2 | |
| LLaVA-STSetting=LLM, Zero-shot=true2025.12 | 21.6 | 8.7 | 6 | 1.2 |