Claim · model-written summary
Against baseline, recall improved 4.1 percentage points.
Not reportedTaxonMaterialGeneCompoundOrgan / stageInteractorInteractor roleSecond interactorSecond interactor roleTreatmentSettingQualifier
Verbatim
e0“Compared to the baseline, YOLO-MECD has improved by 0.2, 4.1, and 3.9 percentage points in detection precision, recall rate, and mAP value, respectively”
Title
YOLO-MECD: Citrus Detection Algorithm Based on YOLOv11
Yue Liao; Lerong Li; Huiqiang Xiao; et al. Agronomy 2025. doi:10.3390/agronomy15030687
Other claims from this paper (15)
- Method / resource Replacing C2PSA with EMA improved citrus detection accuracy and reduced model parameters.
- Method / resource CSPPC reduced parameter count and computational complexity while maintaining mAP.
- Method / resource MPDIoU improved bounding-box accuracy and convergence.
- Method / resource Reducing backbone convolutions improved small-object detection and decreased parameters.
- Method / resource YOLO-MECD precision was 84.4%, 0.2 points above baseline.
- Method / resource YOLO-MECD recall was 73.3%.
- Method / resource YOLO-MECD mAP was 81.6%.
- Method / resource Against baseline, precision improved 0.2 percentage points.
- Method / resource Against baseline, mAP improved 3.9 percentage points.
- Method / resource Model parameters fell from 9,413,574 to 2,297,334, a 75.6% decrease.
- Method / resource Model size fell from 18.2 MB to 4.66 MB, a 74.4% reduction.
- Method / resource mAP exceeded YOLOv8s by 3.8%.
- Method / resource mAP exceeded YOLOv9s by 3.2%.
- Method / resource mAP exceeded YOLOv10s by 5.5%.
- Method / resource The model detected pomelos and kumquats well.