Localization-aware captioning on Simulated Radar Dataset (30-40m)
88.8PrecisionSG-CLIP
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| SG-CLIPalpha=4.02025.11 | 88.8 | 84.8 | 86.7 | |
| SG-CLIPalpha=1.02025.11 | 88.7 | 84.8 | 86.7 | |
| SG-CLIPalpha=16.02025.11 | 87 | 84.4 | 85.6 | |
| Vanilla CLIP2025.11 | 67.3 | 52.9 | 57.7 |