Table 17 Performance comparison Of The YOLOv5 combination with different strategies.
From: PEYOLO a perception efficient network for multiscale surface defects detection
Method | DcPAN | PEHead | RFEM | Params↓ | \({\text{mAP}}_{50}\)↑ | FPS↑ | P↑ | R↑ |
|---|---|---|---|---|---|---|---|---|
Baseline | – | – | – | 2.65 | 72.9 | 184.1 | 0.637 | 0.714 |
√ | – | – | 2.73 | 76.2 | 154.4 | 0.70 | 0.714 | |
– | √ | – | 2.41 | 74.7 | 133.4 | 0.698 | 0.691 | |
– | – | √ | 2.92 | 73.7 | 179.6 | 0.641 | 0.71 |
