One-Shot / Visual Prompt Detection Benchmark
Compare object detection models for security, UAV, traffic monitoring, and public safety operations.
| Model |
Detected |
Recall |
Latency |
Best Use Case |
Status |
| YOLOv11 |
19 / 20 |
95% |
18 ms |
Real-time security / traffic |
BEST |
| Grounding DINO |
20 / 20 |
100% |
220 ms |
Visual prompt search |
HIGH ACCURACY |
| Florence-2 |
17 / 20 |
85% |
180 ms |
Caption + detection |
GOOD |
| Open-Vocabulary Detector |
16 / 20 |
80% |
140 ms |
Unknown object detection |
GOOD |
| Baseline Detector |
11 / 20 |
55% |
40 ms |
Legacy camera analytics |
LOW |