Spatio-temporal Video Grounding on VidSTG (test)
34.67m_vIoUVideoG-DINO
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| VideoG-DINOsupervision=Fully-supervised (F-S)2026.06 | 34.67 | 51.97 | 48.11 | 33.96 | |
| CG-STVGsupervision=Fully-supervised (F-S)2026.06 | 34 | 51.4 | 47.7 | 33.1 | |
| CSDVLsupervision=Fully-supervised (F-S)2026.06 | 33.7 | — | 47.2 | 32.8 | |
| STCATsupervision=Fully-supervised (F-S)2026.06 | 33.14 | 50.82 | 46.2 | 32.58 | |
| TubeDETRsupervision=Fully-supervised (F-S)2026.06 | 30.4 | 48.1 | 42.5 | 28.2 | |
| Video RSVG-ZeroOVsupervision=Zero-shot (Z-S), temporal_strategy=Strategy 12026.06 | 17.71 | 30.16 | 21.12 | 15.26 | |
| E3Msupervision=Zero-shot (Z-S)2026.06 | 16.21 | — | 20.47 | 11.91 | |
| CosPaLsupervision=Weakly-supervised (W-S)2026.06 | 16 | — | 20.1 | 13.1 | |
| STProsupervision=Weakly-supervised (W-S)2026.06 | 15.52 | — | 19.39 | 12.69 | |
| VEMsupervision=Weakly-supervised (W-S)2026.06 | 14.45 | — | 18.57 | 8.76 | |
| ReCLIPsupervision=Zero-shot (Z-S)2026.06 | 14.21 | — | 17.54 | 7.86 | |
| WINNERsupervision=Weakly-supervised (W-S)2026.06 | 11.61 | — | 14.12 | 7.4 | |
| Video RSVG-ZeroOVsupervision=Zero-shot (Z-S), temporal_strategy=Strategy 22026.06 | 11.29 | 28.35 | 12.98 | 7.6 | |
| Qwen-2.5-VLsupervision=Zero-shot (Z-S), temporal_strategy=Strategy 12026.06 | 8.79 | 12.43 | 10.2 | 7.67 | |
| RedCirclesupervision=Zero-shot (Z-S)2026.06 | 8.56 | — | 7.61 | 0.93 | |
| Qwen-2.5-VLsupervision=Zero-shot (Z-S), temporal_strategy=Strategy 22026.06 | 6.08 | 12.91 | 7.07 | 4.58 |