Table 7 Performance comparison of different detection algorithms.
From: An improved YOLOv5n algorithm for detecting surface defects in industrial components
Method | AP (%) | mAP@50 (%) | |||||
|---|---|---|---|---|---|---|---|
CR | IN | PA | PS | RS | SC | ||
SSD | 28 | 73.1 | 83.4 | 73.5 | 54.3 | 33.1 | 60.1 |
Faster R-CNN | 35.7 | 80.6 | 79.8 | 84.6 | 52.6 | 85.3 | 70.5 |
YOLOv3 | 28.3 | 75 | 82.4 | 80.6 | 52.1 | 88.3 | 68.2 |
YOLOv4 | 36.5 | 76.8 | 81.8 | 83.2 | 52.9 | 85.9 | 70.6 |
YOLOv5n | 35.8 | 85.8 | 80.8 | 85.5 | 51.6 | 86.4 | 71 |
YOLOv7 | 35.6 | 83.5 | 81 | 86 | 53.2 | 86.5 | 71.3 |
YOLOV8s | 38.2 | 86.9 | 80.4 | 82.6 | 54 | 88.6 | 72.3 |
YOLOv9-t | 42.3 | 89.5 | 82.1 | 88.3 | 53.1 | 89.1 | 74.1 |
YOLOv10n | 36.4 | 85.1 | 80.6 | 83.5 | 53.4 | 88.5 | 72.4 |
PP-YOLOE-s | 37.8 | 84.2 | 79.6 | 84.4 | 52.4 | 87.2 | 72.1 |
Proposed algorithm | 50.4 | 90.5 | 84.5 | 86.1 | 54.3 | 89.7 | 75.3 |
