Object Detection on COCO standard (2%)
34.7mAPMixPL
Evaluation Results
| Method | Links | |
|---|---|---|
| MixPLDetector=DINO2023.12 | 34.7 | |
| Efficient TeacherDetector type=One-stage anchor-based, Backbone=YOLOv5l, FLOPS=109.59G2023.02 | 28.7 | |
| Unbiased Teacher v2Thresholding Strategy=Fully automatic2022.12 | 28.37 | |
| MixPLDetector=FCOS2023.12 | 27.8 | |
| PseCoDetector type=Two-stage anchor-based, FLOPS=202.31G2023.02 | 27.77 | |
| PseCoThresholding Strategy=Fully automatic2022.12 | 27.77 | |
| Dense TeacherDetector type=One-stage anchor-free, FLOPS=200.59G2023.02 | 27.2 | |
| Dense TeacherDetector=FCOS2023.12 | 27.2 | |
| Efficient TeacherDetector type=One-stage anchor-based, FLOPS=169.61G2023.02 | 27.15 | |
| Unbiased Teacher v2Detector type=One-stage anchor-free, FLOPS=200.59G2023.02 | 26.03 | |
| DSLDetector=FCOS2023.12 | 25.2 | |
| DSLDetector type=One-stage anchor-free, FLOPS=200.59G2023.02 | 25.19 | |
| ASTODThresholding Strategy=Fully automatic, Iterations=32022.12 | 24.85 | |
| Unbiased Teacher2022.03 | 24.3 | |
| Unbiased TeacherDetector type=Two-stage anchor-based, FLOPS=204.13G2023.02 | 24.3 | |
| Unbiased TeacherThresholding Strategy=Manual empirical search2022.12 | 24.3 | |
| Omni-DETR2022.03 | 23.2 | |
| Omni-DETRThresholding Strategy=Manual empirical search2022.12 | 23.2 | |
| Unbiased Teacher (re-implemented)Detector type=One-stage anchor-based, FLOPS=169.61G2023.02 | 22.72 | |
| Instant TeachingDetector type=Two-stage anchor-based, FLOPS=202.31G2023.02 | 22.45 | |
| Instant TeachingThresholding Strategy=Manual empirical search2022.12 | 22.45 | |
| ISMTThresholding Strategy=Manual empirical search2022.12 | 22.43 | |
| Humber teacherDetector type=Two-stage anchor-based, FLOPS=202.31G2023.02 | 21.72 | |
| Humble TeacherThresholding Strategy=Fully automatic2022.12 | 21.72 | |
| Humble Teacher2022.03 | 21.7 | |
| STAC+VL-PLM2022.07 | 21.2 | |
| Supervised+VL-PLM2022.07 | 18.6 | |
| STAC2022.03 | 18.3 | |
| STACDetector type=Two-stage anchor-based, FLOPS=202.31G2023.02 | 18.25 | |
| STACThresholding Strategy=Manual empirical search2022.12 | 18.25 | |
| STAC2022.07 | 18.25 | |
| Supervised†Thresholding Strategy=Baseline / Teacher2022.12 | 16.67 | |
| Faster R-CNNImplementation=Re-implementation2022.03 | 14.9 | |
| Supervised+PLs2022.07 | 14.88 | |
| Deformable DETRImplementation=Re-implementation2022.03 | 14.7 | |
| CSDThresholding Strategy=Fully automatic2022.12 | 13.93 | |
| Supervised (One-stage anchor-based)Detector type=One-stage anchor-based, FLOPS=169.61G2023.02 | 13.12 | |
| Faster R-CNN2022.03 | 12.7 | |
| Supervised (Two-stage)Detector type=Two-stage anchor-based, FLOPS=202.31G2023.02 | 12.7 | |
| SupervisedThresholding Strategy=Baseline2022.12 | 12.7 | |
| Supervised2022.07 | 12.7 | |
| Supervised (One-stage anchor-free)Detector type=One-stage anchor-free, FLOPS=200.59G2023.02 | 11.71 |