Table 4 Ablation experiment.
From: Research on steel surface defect classification method based on deep learning
Model | K-means | SE | BiFPN | mAP50 | FPS |
|---|---|---|---|---|---|
YOLOv5l | Â | Â | Â | 0.757 | 71 |
√ |  |  | 0.767 | 71 | |
√ | √ |  | 0.771 | 71 | |
√ |  | √ | 0.771 | 70 | |
√ | √ | √ | 0.799 | 70 |
From: Research on steel surface defect classification method based on deep learning
Model | K-means | SE | BiFPN | mAP50 | FPS |
|---|---|---|---|---|---|
YOLOv5l | Â | Â | Â | 0.757 | 71 |
√ |  |  | 0.767 | 71 | |
√ | √ |  | 0.771 | 71 | |
√ |  | √ | 0.771 | 70 | |
√ | √ | √ | 0.799 | 70 |
