Object Detection on TruckDrive (test)
49.7COCO mAPTelescope
Evaluation results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| TelescopeBackbone=SAM3, Detection Head=Deformable DETR, Training Protocol=Fine-tuned on TruckDrive (12 epochs), Components=de-noising training, TeleScope re-sampling layer2026.04 | 49.7 | 60.8 | 50.7 | 33.5 | 32.6 | 80.1 | 49.4 | |
| RVSABackbone=ImageNet pre-trained, Training Protocol=Fine-tuned on TruckDrive (12 epochs)2026.04 | 32.5 | 50.2 | 29.8 | 23.3 | 18.3 | 48.8 | 35.1 | |
| RFLABackbone=ImageNet pre-trained, Training Protocol=Fine-tuned on TruckDrive (12 epochs)2026.04 | 30.6 | 50.1 | 32 | 23.9 | 18.5 | 51.2 | 31.7 | |
| UniverseNetBackbone=ImageNet pre-trained, Training Protocol=Fine-tuned on TruckDrive (12 epochs)2026.04 | 30.5 | 51.8 | 33.4 | 23.6 | 14.5 | 47.4 | 31.8 | |
| Grounding DINOBackbone=ImageNet pre-trained, Training Protocol=Fine-tuned on TruckDrive (12 epochs)2026.04 | 28.6 | 37.6 | 26.2 | 14.7 | 15.6 | 41.7 | 29.6 | |
| YOLO11xBackbone=ImageNet pre-trained, Training Protocol=Fine-tuned on TruckDrive (12 epochs)2026.04 | 26.6 | 42.1 | 21.8 | 13.4 | 11.7 | 51 | 19.5 | |
| QueryDetBackbone=ImageNet pre-trained, Training Protocol=Fine-tuned on TruckDrive (12 epochs)2026.04 | 24.8 | 44.9 | 28.6 | 19.9 | 9.4 | 41.5 | 25.7 | |
| DINOBackbone=ImageNet pre-trained, Training Protocol=Fine-tuned on TruckDrive (12 epochs)2026.04 | 22.2 | 33.5 | 23.9 | 18.9 | 17.9 | 37.1 | 22.6 | |
| DETRBackbone=ImageNet pre-trained, Training Protocol=Fine-tuned on TruckDrive (12 epochs)2026.04 | 16.6 | 39.6 | 17.8 | 8.1 | 7.2 | 33.5 | 14.7 | |
| FOVEABackbone=ImageNet pre-trained, Training Protocol=Fine-tuned on TruckDrive (12 epochs)2026.04 | 11.3 | 16.9 | 8.6 | 0.8 | 0.5 | 18.9 | 11.5 |