Table 11 Evaluation metric results of the comparative experiment on the AL10-DET dataset.
From: Metal surface defect detection using SLF-YOLO enhanced YOLOv8 model
Metric | SSD | Faster-RCNN | RTDETR | v3-tiny | v4 | v5s | v6s | v7 | v8s | Ours |
|---|---|---|---|---|---|---|---|---|---|---|
P(%) | 78.4 | 76.0 | 84.9 | 71.0 | 84.6 | 86.5 | 85.1 | 77.5 | 83.4 | 85.7 |
R(%) | 82.0 | 71.4 | 80.3 | 72.5 | 55.0 | 76.5 | 77.1 | 72.6 | 80.1 | 81.4 |
mAP50(%) | 80.6 | 73.6 | 82.1 | 75.2 | 65.8 | 81.8 | 81.8 | 75.9 | 83.0 | 86.8 |
mAP50:95(%) | 48.6 | 39.8 | 63.1 | 49.7 | 40.5 | 64.7 | 63.8 | 57.2 | 65.3 | 68.9 |
Param(M) | 23.3 | 41.5 | 41.9 | 12.1 | 64.0 | 9.1 | 18.5 | 36.5 | 11.12 | 9.65 |
Flops(G) | 36.0 | 108.0 | 125.6 | 19.1 | 70.0 | 23.8 | 45.2 | 103.2 | 28.4 | 24.6 |
