Table 4 Detection evaluation results (%) of each model against the proposed DCBS-YOLO.
From: Explainable hybrid AI CAD framework for advanced prediction of steel surface defects
Methods | Cr | In | Pa | Ps | Rs | Sc | mAP@0.5 (%) |
|---|---|---|---|---|---|---|---|
YOLOv5n | 17.2 | 74.7 | 95.8 | 91.8 | 48.0 | 75.5 | 67.1 |
YOLOv5s | 21.8 | 79.2 | 95.5 | 90.4 | 54.3 | 82.2 | 70.5 |
YOLOv5m | 33.2 | 81.7 | 96.4 | 90.7 | 56.5 | 88.2 | 74.4 |
YOLOv5l | 26.0 | 79.5 | 97.3 | 90.9 | 55.0 | 88.8 | 72.9 |
YOLOv5x | 28.8 | 78.5 | 96.8 | 85.9 | 64.5 | 88.8 | 73.8 |
YOLOv8n | 23.9 | 79.9 | 97.0 | 95.0 | 53.5 | 85.1 | 72.4 |
YOLOv8s | 22.2 | 72.8 | 95.9 | 93.4 | 55.1 | 88.9 | 71.3 |
YOLOv8m | 23.2 | 81.0 | 97.5 | 95.1 | 56.8 | 79.8 | 72.2 |
YOLOv8l | 24.1 | 80.1 | 97.5 | 87.1 | 56.0 | 85.6 | 71.7 |
YOLOv8x | 31.6 | 77.3 | 96.9 | 90.0 | 57.4 | 89.2 | 73.7 |
YOLOv9t | 21.1 | 79.7 | 96.5 | 94.6 | 52.5 | 82.9 | 71.2 |
YOLOv9s | 28.6 | 79.1 | 97.1 | 93.7 | 59.8 | 91.5 | 74.9 |
YOLOv9m | 33.7 | 81.0 | 95.7 | 93.2 | 59.4 | 85.1 | 74.6 |
YOLOv9c | 33.5 | 81.2 | 97.3 | 94.6 | 61.2 | 82.1 | 75.0 |
YOLOv9e | 29.8 | 81.3 | 97.8 | 85.8 | 55.8 | 85.8 | 72.7 |
The Proposed DCBS-YOLO | 25.8 | 81.6 | 98.4 | 95.3 | 60.8 | 92.5 | 75.7 |
