Aerial Visual Grounding on AerialSense (test)
51.09AccuracyGemini-2.0
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Gemini-2.0Semantic Enhancement=true, Spatial Position Enhancement=true, Spatial Relationship Enhancement=true2025.12 | 51.09 | 46.22 | |
| GPT-4.1Semantic Enhancement=true, Spatial Position Enhancement=true, Spatial Relationship Enhancement=true2025.12 | 49.28 | 44.87 | |
| GPT-4oSemantic Enhancement=true, Spatial Position Enhancement=true, Spatial Relationship Enhancement=true2025.12 | 45.05 | 41.26 | |
| Claude-3.7Semantic Enhancement=true, Spatial Position Enhancement=true, Spatial Relationship Enhancement=true2025.12 | 44.93 | 41.08 | |
| Qwen2.5-MaxSemantic Enhancement=true, Spatial Position Enhancement=true, Spatial Relationship Enhancement=true2025.12 | 43.87 | 39.87 | |
| LLaVA-OneVision-7BSemantic Enhancement=true, Spatial Position Enhancement=true, Spatial Relationship Enhancement=true2025.12 | 38.65 | 35.69 | |
| Qwen2-VL-7BSemantic Enhancement=true, Spatial Position Enhancement=true, Spatial Relationship Enhancement=true2025.12 | 37.94 | 34.79 | |
| SAILVL-8BSemantic Enhancement=true, Spatial Position Enhancement=true, Spatial Relationship Enhancement=true2025.12 | 34.23 | 31.82 | |
| InternVL3-14BSemantic Enhancement=true, Spatial Position Enhancement=true, Spatial Relationship Enhancement=true2025.12 | 26.59 | 24.48 | |
| LLaVA-1.5-7BSemantic Enhancement=true, Spatial Position Enhancement=true, Spatial Relationship Enhancement=true2025.12 | 22.88 | 21.61 | |
| Qwen2-VL-7BSemantic Enhancement=false, Spatial Position Enhancement=false, Spatial Relationship Enhancement=false2025.12 | 17.94 | 19.43 | |
| InternVL3-8BSemantic Enhancement=true, Spatial Position Enhancement=true, Spatial Relationship Enhancement=true2025.12 | 14.21 | 14.5 | |
| InternVL3-14BSemantic Enhancement=false, Spatial Position Enhancement=false, Spatial Relationship Enhancement=false2025.12 | 11.62 | 18.36 | |
| Gemini-2.0Semantic Enhancement=false, Spatial Position Enhancement=false, Spatial Relationship Enhancement=false2025.12 | 10.17 | 14.99 | |
| Claude-3.7Semantic Enhancement=false, Spatial Position Enhancement=false, Spatial Relationship Enhancement=false2025.12 | 8.8 | 12.94 | |
| LLaVA-OneVision-7BSemantic Enhancement=false, Spatial Position Enhancement=false, Spatial Relationship Enhancement=false2025.12 | 8.63 | 14.26 | |
| LLaVA-1.5-7BSemantic Enhancement=false, Spatial Position Enhancement=false, Spatial Relationship Enhancement=false2025.12 | 7.6 | 13.57 | |
| SAILVL-8BSemantic Enhancement=false, Spatial Position Enhancement=false, Spatial Relationship Enhancement=false2025.12 | 6.15 | 10.93 | |
| GPT-4.1Semantic Enhancement=false, Spatial Position Enhancement=false, Spatial Relationship Enhancement=false2025.12 | 3.6 | 11.91 | |
| InternVL3-8BSemantic Enhancement=false, Spatial Position Enhancement=false, Spatial Relationship Enhancement=false2025.12 | 2.4 | 7.38 | |
| GPT-4oSemantic Enhancement=false, Spatial Position Enhancement=false, Spatial Relationship Enhancement=false2025.12 | 2.04 | 3.69 | |
| Qwen2.5-MaxSemantic Enhancement=false, Spatial Position Enhancement=false, Spatial Relationship Enhancement=false2025.12 | 1.22 | 7.17 |