Localization-aware captioning on Simulated Radar Dataset (20-30m)
87.5PrecisionSG-CLIP
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| SG-CLIPalpha=4.02025.11 | 87.5 | 83.4 | 85.2 | |
| SG-CLIPalpha=1.02025.11 | 86.3 | 85.8 | 86.1 | |
| SG-CLIPalpha=16.02025.11 | 84.9 | 81.6 | 82.9 | |
| Vanilla CLIP2025.11 | 74.2 | 66.9 | 69.3 |