Table 2 Ablation experiments for each part of the model.
From: A photovoltaic cell defect detection model capable of topological knowledge extraction
Modul | p | R | mAP@0.5 | Modul | p | R | mAP@0.5 |
---|---|---|---|---|---|---|---|
YOLO-V5 | 0.875 | 0.881 | 0.905 | YOLO-V7[29] | 0.879 | 0.824 | 0.880 |
PicoDet[26] | 0.871 | 0.841 | 0.775 | YOLO-V8 | 0.890 | 0.872 | 0.886 |
YOLO-X[27] | 0.881 | 0.865 | 0.896 | Deformable-DETR[30] | \ | 0.755 | 0.842 |
Cascade-RCNN | 0.912 | 0.881 | 0.893 | Faster-RCNN | 0.818 | 0.805 | 0.824 |
Sparse-RCNN[28] | 0.841 | 0.817 | 0.840 | Ours | 0.891 | 0.894 | 0.920 |